Spaces:
Sleeping
Sleeping
black .
Browse files- app.py +314 -159
- classifiers.py +72 -66
- prompts.py +1 -1
- utils.py +61 -48
app.py
CHANGED
@@ -20,12 +20,13 @@ from prompts import (
|
|
20 |
CATEGORY_SUGGESTION_PROMPT,
|
21 |
ADDITIONAL_CATEGORY_PROMPT,
|
22 |
VALIDATION_ANALYSIS_PROMPT,
|
23 |
-
CATEGORY_IMPROVEMENT_PROMPT
|
24 |
)
|
25 |
|
26 |
# Configure logging
|
27 |
-
logging.basicConfig(
|
28 |
-
|
|
|
29 |
|
30 |
# Initialize API key from environment variable
|
31 |
OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY", "")
|
@@ -39,22 +40,23 @@ if OPENAI_API_KEY:
|
|
39 |
except Exception as e:
|
40 |
logging.error(f"Failed to initialize OpenAI client: {str(e)}")
|
41 |
|
|
|
42 |
def update_api_key(api_key):
|
43 |
"""Update the OpenAI API key"""
|
44 |
global OPENAI_API_KEY, client
|
45 |
-
|
46 |
if not api_key:
|
47 |
return "API Key cannot be empty"
|
48 |
-
|
49 |
OPENAI_API_KEY = api_key
|
50 |
-
|
51 |
try:
|
52 |
client = OpenAI(api_key=api_key)
|
53 |
# Test the connection with a simple request
|
54 |
response = client.chat.completions.create(
|
55 |
model="gpt-3.5-turbo",
|
56 |
messages=[{"role": "user", "content": "test"}],
|
57 |
-
max_tokens=5
|
58 |
)
|
59 |
return f"API Key updated and verified successfully"
|
60 |
except Exception as e:
|
@@ -62,41 +64,45 @@ def update_api_key(api_key):
|
|
62 |
logging.error(f"API key update failed: {error_msg}")
|
63 |
return f"Failed to update API Key: {error_msg}"
|
64 |
|
|
|
65 |
def process_file(file, text_columns, categories, classifier_type, show_explanations):
|
66 |
"""Process the uploaded file and classify text data"""
|
67 |
# Initialize result_df and validation_report
|
68 |
result_df = None
|
69 |
validation_report = None
|
70 |
-
|
71 |
try:
|
72 |
# Load data from file
|
73 |
if isinstance(file, str):
|
74 |
df = load_data(file)
|
75 |
else:
|
76 |
df = load_data(file.name)
|
77 |
-
|
78 |
if not text_columns:
|
79 |
return None, "Please select at least one text column"
|
80 |
-
|
81 |
# Check if all selected columns exist
|
82 |
missing_columns = [col for col in text_columns if col not in df.columns]
|
83 |
if missing_columns:
|
84 |
-
return
|
85 |
-
|
|
|
|
|
|
|
86 |
# Combine text from selected columns
|
87 |
texts = []
|
88 |
for _, row in df.iterrows():
|
89 |
combined_text = " ".join(str(row[col]) for col in text_columns)
|
90 |
texts.append(combined_text)
|
91 |
-
|
92 |
# Parse categories if provided
|
93 |
category_list = []
|
94 |
if categories:
|
95 |
category_list = [cat.strip() for cat in categories.split(",")]
|
96 |
-
|
97 |
# Select classifier based on data size and user choice
|
98 |
num_texts = len(texts)
|
99 |
-
|
100 |
# If no specific model is chosen, select the most appropriate one
|
101 |
if classifier_type == "auto":
|
102 |
if num_texts <= 500:
|
@@ -107,30 +113,36 @@ def process_file(file, text_columns, categories, classifier_type, show_explanati
|
|
107 |
classifier_type = "hybrid"
|
108 |
else:
|
109 |
classifier_type = "tfidf"
|
110 |
-
|
111 |
# Initialize appropriate classifier
|
112 |
if classifier_type == "tfidf":
|
113 |
classifier = TFIDFClassifier()
|
114 |
results = classifier.classify(texts, category_list)
|
115 |
elif classifier_type in ["gpt35", "gpt4"]:
|
116 |
if client is None:
|
117 |
-
return
|
|
|
|
|
|
|
118 |
model = "gpt-3.5-turbo" if classifier_type == "gpt35" else "gpt-4"
|
119 |
classifier = LLMClassifier(client=client, model=model)
|
120 |
results = classifier.classify(texts, category_list)
|
121 |
else: # hybrid
|
122 |
if client is None:
|
123 |
-
return
|
|
|
|
|
|
|
124 |
# First pass with TF-IDF
|
125 |
tfidf_classifier = TFIDFClassifier()
|
126 |
tfidf_results = tfidf_classifier.classify(texts, category_list)
|
127 |
-
|
128 |
# Second pass with LLM for low confidence results
|
129 |
llm_classifier = LLMClassifier(client=client, model="gpt-3.5-turbo")
|
130 |
results = []
|
131 |
low_confidence_texts = []
|
132 |
low_confidence_indices = []
|
133 |
-
|
134 |
for i, (text, tfidf_result) in enumerate(zip(texts, tfidf_results)):
|
135 |
if tfidf_result["confidence"] < 70: # If confidence is below 70%
|
136 |
low_confidence_texts.append(text)
|
@@ -138,91 +150,97 @@ def process_file(file, text_columns, categories, classifier_type, show_explanati
|
|
138 |
results.append(None) # Placeholder
|
139 |
else:
|
140 |
results.append(tfidf_result)
|
141 |
-
|
142 |
if low_confidence_texts:
|
143 |
-
llm_results = llm_classifier.classify(
|
|
|
|
|
144 |
for idx, llm_result in zip(low_confidence_indices, llm_results):
|
145 |
results[idx] = llm_result
|
146 |
-
|
147 |
# Create results dataframe
|
148 |
result_df = df.copy()
|
149 |
result_df["Category"] = [r["category"] for r in results]
|
150 |
result_df["Confidence"] = [r["confidence"] for r in results]
|
151 |
-
|
152 |
if show_explanations:
|
153 |
result_df["Explanation"] = [r["explanation"] for r in results]
|
154 |
-
|
155 |
# Validate results using LLM
|
156 |
validation_report = validate_results(result_df, text_columns, client)
|
157 |
-
|
158 |
return result_df, validation_report
|
159 |
-
|
160 |
except Exception as e:
|
161 |
error_traceback = traceback.format_exc()
|
162 |
return None, f"Error: {str(e)}\n{error_traceback}"
|
163 |
|
|
|
164 |
def export_results(df, format_type):
|
165 |
"""Export results to a file and return the file path for download"""
|
166 |
if df is None:
|
167 |
return None
|
168 |
-
|
169 |
# Create a temporary file
|
170 |
import tempfile
|
171 |
import os
|
172 |
-
|
173 |
# Create a temporary directory if it doesn't exist
|
174 |
temp_dir = "temp_exports"
|
175 |
os.makedirs(temp_dir, exist_ok=True)
|
176 |
-
|
177 |
# Generate a unique filename
|
178 |
timestamp = time.strftime("%Y%m%d-%H%M%S")
|
179 |
filename = f"classification_results_{timestamp}"
|
180 |
-
|
181 |
if format_type == "excel":
|
182 |
file_path = os.path.join(temp_dir, f"{filename}.xlsx")
|
183 |
df.to_excel(file_path, index=False)
|
184 |
else:
|
185 |
file_path = os.path.join(temp_dir, f"{filename}.csv")
|
186 |
df.to_csv(file_path, index=False)
|
187 |
-
|
188 |
return file_path
|
189 |
|
|
|
190 |
# Create Gradio interface
|
191 |
with gr.Blocks(title="Text Classification System") as demo:
|
192 |
gr.Markdown("# Text Classification System")
|
193 |
gr.Markdown("Upload your data file (Excel/CSV) and classify text using AI")
|
194 |
-
|
195 |
with gr.Tab("Setup"):
|
196 |
api_key_input = gr.Textbox(
|
197 |
label="OpenAI API Key",
|
198 |
placeholder="Enter your API key here",
|
199 |
type="password",
|
200 |
-
value=OPENAI_API_KEY
|
201 |
)
|
202 |
api_key_button = gr.Button("Update API Key")
|
203 |
api_key_message = gr.Textbox(label="Status", interactive=False)
|
204 |
-
|
205 |
# Display current API status
|
206 |
-
api_status =
|
|
|
|
|
207 |
gr.Markdown(f"**Current API Status**: {api_status}")
|
208 |
-
|
209 |
-
api_key_button.click(
|
210 |
-
|
|
|
|
|
211 |
with gr.Tab("Classify Data"):
|
212 |
with gr.Column():
|
213 |
file_input = gr.File(label="Upload Excel/CSV File")
|
214 |
-
|
215 |
# Variable to store available columns
|
216 |
available_columns = gr.State([])
|
217 |
-
|
218 |
# Button to load file and suggest categories
|
219 |
load_categories_button = gr.Button("Load File")
|
220 |
-
|
221 |
# Display original dataframe
|
222 |
original_df = gr.Dataframe(
|
223 |
-
label="Original Data",
|
224 |
-
interactive=False,
|
225 |
-
visible=False
|
226 |
)
|
227 |
|
228 |
with gr.Row():
|
@@ -232,31 +250,29 @@ with gr.Blocks(title="Text Classification System") as demo:
|
|
232 |
choices=[],
|
233 |
value=[],
|
234 |
interactive=True,
|
235 |
-
visible=False
|
236 |
)
|
237 |
|
238 |
new_category = gr.Textbox(
|
239 |
label="Add New Category",
|
240 |
placeholder="Enter a new category name",
|
241 |
-
visible=False
|
242 |
)
|
243 |
with gr.Row():
|
244 |
add_category_button = gr.Button("Add Category", visible=False)
|
245 |
-
suggest_category_button = gr.Button(
|
246 |
-
|
|
|
247 |
|
248 |
# Original categories input (hidden)
|
249 |
-
categories = gr.Textbox(
|
250 |
-
|
251 |
-
)
|
252 |
-
|
253 |
-
|
254 |
with gr.Column():
|
255 |
text_column = gr.CheckboxGroup(
|
256 |
-
label="Select Text Columns",
|
257 |
-
choices=[],
|
258 |
interactive=True,
|
259 |
-
visible=False
|
260 |
)
|
261 |
|
262 |
classifier_type = gr.Dropdown(
|
@@ -264,18 +280,20 @@ with gr.Blocks(title="Text Classification System") as demo:
|
|
264 |
("TF-IDF (Rapide, <1000 lignes)", "tfidf"),
|
265 |
("LLM GPT-3.5 (Fiable, <1000 lignes)", "gpt35"),
|
266 |
("LLM GPT-4 (Très fiable, <500 lignes)", "gpt4"),
|
267 |
-
("TF-IDF + LLM (Hybride, >1000 lignes)", "hybrid")
|
268 |
],
|
269 |
label="Modèle de classification",
|
270 |
value="gpt35",
|
271 |
-
visible=False
|
|
|
|
|
|
|
272 |
)
|
273 |
-
|
274 |
-
|
275 |
process_button = gr.Button("Process and Classify", visible=False)
|
276 |
|
277 |
results_df = gr.Dataframe(interactive=True, visible=False)
|
278 |
-
|
279 |
# Create containers for visualization and validation report
|
280 |
with gr.Row(visible=False) as results_row:
|
281 |
with gr.Column():
|
@@ -284,161 +302,251 @@ with gr.Blocks(title="Text Classification System") as demo:
|
|
284 |
csv_download = gr.File(label="Download CSV", visible=False)
|
285 |
excel_download = gr.File(label="Download Excel", visible=False)
|
286 |
with gr.Column():
|
287 |
-
validation_output = gr.Textbox(
|
288 |
-
|
|
|
|
|
|
|
|
|
289 |
|
290 |
# Function to load file and suggest categories
|
291 |
def load_file_and_suggest_categories(file):
|
292 |
if not file:
|
293 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
294 |
try:
|
295 |
df = load_data(file.name)
|
296 |
columns = list(df.columns)
|
297 |
-
|
298 |
# Analyze columns to suggest text columns
|
299 |
suggested_text_columns = []
|
300 |
for col in columns:
|
301 |
# Check if column contains text data
|
302 |
-
if df[col].dtype ==
|
303 |
# Check if column contains mostly text (not just numbers or dates)
|
304 |
sample = df[col].head(100).dropna()
|
305 |
if len(sample) > 0:
|
306 |
# Check if most values contain spaces (indicating text)
|
307 |
-
text_ratio = sum(
|
308 |
-
|
|
|
|
|
|
|
|
|
309 |
suggested_text_columns.append(col)
|
310 |
-
|
311 |
# If no columns were suggested, use all object columns
|
312 |
if not suggested_text_columns:
|
313 |
-
suggested_text_columns = [
|
314 |
-
|
|
|
|
|
315 |
# Get a sample of text for category suggestion
|
316 |
sample_texts = []
|
317 |
for col in suggested_text_columns:
|
318 |
sample_texts.extend(df[col].head(5).tolist())
|
319 |
-
|
320 |
# Use LLM to suggest categories
|
321 |
if client:
|
322 |
-
prompt = CATEGORY_SUGGESTION_PROMPT.format(
|
|
|
|
|
323 |
try:
|
324 |
response = client.chat.completions.create(
|
325 |
model="gpt-3.5-turbo",
|
326 |
messages=[{"role": "user", "content": prompt}],
|
327 |
temperature=0,
|
328 |
-
max_tokens=100
|
329 |
)
|
330 |
-
suggested_cats = [
|
|
|
|
|
|
|
|
|
|
|
331 |
except:
|
332 |
-
suggested_cats = [
|
|
|
|
|
|
|
|
|
|
|
|
|
333 |
else:
|
334 |
-
suggested_cats = [
|
335 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
336 |
return (
|
337 |
-
columns,
|
338 |
-
gr.CheckboxGroup(choices=columns, value=suggested_text_columns),
|
339 |
-
gr.CheckboxGroup(
|
|
|
|
|
340 |
gr.Textbox(visible=True),
|
341 |
gr.Button(visible=True),
|
342 |
gr.Button(visible=True),
|
343 |
-
gr.CheckboxGroup(
|
|
|
|
|
344 |
gr.Dropdown(visible=True),
|
345 |
gr.Checkbox(visible=True),
|
346 |
gr.Button(visible=True),
|
347 |
-
gr.Dataframe(value=df, visible=True)
|
348 |
)
|
349 |
except Exception as e:
|
350 |
-
return
|
351 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
352 |
# Function to add a new category
|
353 |
def add_new_category(current_categories, new_category):
|
354 |
if not new_category or new_category.strip() == "":
|
355 |
return current_categories
|
356 |
new_categories = current_categories + [new_category.strip()]
|
357 |
return gr.CheckboxGroup(choices=new_categories, value=new_categories)
|
358 |
-
|
359 |
# Function to update categories textbox
|
360 |
def update_categories_textbox(selected_categories):
|
361 |
return ", ".join(selected_categories)
|
362 |
-
|
363 |
# Function to show results after processing
|
364 |
def show_results(df, validation_report):
|
365 |
"""Show the results after processing"""
|
366 |
if df is None:
|
367 |
-
return
|
368 |
-
|
|
|
|
|
|
|
|
|
|
|
369 |
# Export to both formats
|
370 |
csv_path = export_results(df, "csv")
|
371 |
excel_path = export_results(df, "excel")
|
372 |
-
|
373 |
-
return
|
374 |
-
|
|
|
|
|
|
|
|
|
|
|
375 |
# Function to suggest a new category
|
376 |
def suggest_new_category(file, current_categories, text_columns):
|
377 |
if not file or not text_columns:
|
378 |
-
return gr.CheckboxGroup(
|
379 |
-
|
|
|
|
|
380 |
try:
|
381 |
df = load_data(file.name)
|
382 |
-
|
383 |
# Get sample texts from selected columns
|
384 |
sample_texts = []
|
385 |
for col in text_columns:
|
386 |
sample_texts.extend(df[col].head(5).tolist())
|
387 |
-
|
388 |
if client:
|
389 |
prompt = ADDITIONAL_CATEGORY_PROMPT.format(
|
390 |
existing_categories=", ".join(current_categories),
|
391 |
-
sample_texts="\n---\n".join(sample_texts[:10])
|
392 |
)
|
393 |
try:
|
394 |
response = client.chat.completions.create(
|
395 |
model="gpt-3.5-turbo",
|
396 |
messages=[{"role": "user", "content": prompt}],
|
397 |
temperature=0,
|
398 |
-
max_tokens=50
|
399 |
)
|
400 |
new_cat = response.choices[0].message.content.strip()
|
401 |
if new_cat and new_cat not in current_categories:
|
402 |
current_categories.append(new_cat)
|
403 |
except:
|
404 |
pass
|
405 |
-
|
406 |
-
return gr.CheckboxGroup(
|
|
|
|
|
407 |
except Exception as e:
|
408 |
-
return gr.CheckboxGroup(
|
409 |
-
|
|
|
|
|
410 |
# Function to handle export and show download button
|
411 |
def handle_export(df, format_type):
|
412 |
if df is None:
|
413 |
return gr.File(visible=False)
|
414 |
file_path = export_results(df, format_type)
|
415 |
return gr.File(value=file_path, visible=True)
|
416 |
-
|
417 |
# Function to improve classification based on validation report
|
418 |
-
def improve_classification(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
419 |
"""Improve classification based on validation report"""
|
420 |
if df is None or not validation_report:
|
421 |
-
return
|
422 |
-
|
|
|
|
|
|
|
|
|
|
|
423 |
try:
|
424 |
# Extract insights from validation report
|
425 |
if client:
|
426 |
prompt = VALIDATION_ANALYSIS_PROMPT.format(
|
427 |
validation_report=validation_report,
|
428 |
-
current_categories=categories
|
429 |
)
|
430 |
try:
|
431 |
response = client.chat.completions.create(
|
432 |
model="gpt-4",
|
433 |
messages=[{"role": "user", "content": prompt}],
|
434 |
temperature=0,
|
435 |
-
max_tokens=300
|
|
|
|
|
|
|
436 |
)
|
437 |
-
|
438 |
-
|
439 |
# Get current categories
|
440 |
-
current_categories = [
|
441 |
-
|
|
|
|
|
442 |
# If new categories are needed, suggest them based on the data
|
443 |
if improvements.get("new_categories_needed", False):
|
444 |
# Get sample texts for category suggestion
|
@@ -449,51 +557,84 @@ with gr.Blocks(title="Text Classification System") as demo:
|
|
449 |
else:
|
450 |
temp_df = load_data(file.name)
|
451 |
sample_texts.extend(temp_df[col].head(10).tolist())
|
452 |
-
|
453 |
category_prompt = CATEGORY_IMPROVEMENT_PROMPT.format(
|
454 |
current_categories=", ".join(current_categories),
|
455 |
-
analysis=improvements.get(
|
456 |
-
sample_texts="\n---\n".join(sample_texts[:10])
|
457 |
)
|
458 |
-
|
459 |
category_response = client.chat.completions.create(
|
460 |
model="gpt-4",
|
461 |
messages=[{"role": "user", "content": category_prompt}],
|
462 |
temperature=0,
|
463 |
-
max_tokens=100
|
464 |
)
|
465 |
-
|
466 |
-
new_categories = [
|
|
|
|
|
|
|
|
|
|
|
467 |
# Combine current and new categories
|
468 |
all_categories = current_categories + new_categories
|
469 |
categories = ",".join(all_categories)
|
470 |
-
|
471 |
# Process with improved parameters
|
472 |
improved_df, new_validation = process_file(
|
473 |
file,
|
474 |
text_columns,
|
475 |
categories,
|
476 |
classifier_type,
|
477 |
-
show_explanations
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
478 |
)
|
479 |
-
|
480 |
-
return improved_df, new_validation, gr.Button(visible=True), gr.CheckboxGroup(choices=all_categories, value=all_categories)
|
481 |
except Exception as e:
|
482 |
print(f"Error in improvement process: {str(e)}")
|
483 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
484 |
else:
|
485 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
486 |
except Exception as e:
|
487 |
print(f"Error in improvement process: {str(e)}")
|
488 |
-
return
|
489 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
490 |
# Connect functions
|
491 |
load_categories_button.click(
|
492 |
load_file_and_suggest_categories,
|
493 |
inputs=[file_input],
|
494 |
outputs=[
|
495 |
-
available_columns,
|
496 |
-
text_column,
|
497 |
suggested_categories,
|
498 |
new_category,
|
499 |
add_category_button,
|
@@ -502,74 +643,88 @@ with gr.Blocks(title="Text Classification System") as demo:
|
|
502 |
classifier_type,
|
503 |
show_explanations,
|
504 |
process_button,
|
505 |
-
original_df
|
506 |
-
]
|
507 |
)
|
508 |
-
|
509 |
add_category_button.click(
|
510 |
add_new_category,
|
511 |
inputs=[suggested_categories, new_category],
|
512 |
-
outputs=[suggested_categories]
|
513 |
)
|
514 |
-
|
515 |
suggested_categories.change(
|
516 |
update_categories_textbox,
|
517 |
inputs=[suggested_categories],
|
518 |
-
outputs=[categories]
|
519 |
)
|
520 |
-
|
521 |
suggest_category_button.click(
|
522 |
suggest_new_category,
|
523 |
inputs=[file_input, suggested_categories, text_column],
|
524 |
-
outputs=[suggested_categories]
|
525 |
)
|
526 |
-
|
527 |
process_button.click(
|
528 |
-
lambda: gr.Dataframe(visible=True),
|
529 |
-
inputs=[],
|
530 |
-
outputs=[results_df]
|
531 |
).then(
|
532 |
process_file,
|
533 |
-
inputs=[
|
534 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
535 |
).then(
|
536 |
show_results,
|
537 |
inputs=[results_df, validation_output],
|
538 |
-
outputs=[results_row, csv_download, excel_download, results_df]
|
539 |
).then(
|
540 |
-
visualize_results,
|
541 |
-
inputs=[results_df, text_column],
|
542 |
-
outputs=[visualization]
|
543 |
).then(
|
544 |
-
lambda x: gr.Button(visible=True),
|
545 |
-
inputs=[],
|
546 |
-
outputs=[improve_button]
|
547 |
)
|
548 |
-
|
549 |
improve_button.click(
|
550 |
improve_classification,
|
551 |
-
inputs=[
|
552 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
553 |
).then(
|
554 |
show_results,
|
555 |
inputs=[results_df, validation_output],
|
556 |
-
outputs=[results_row, csv_download, excel_download, results_df]
|
557 |
).then(
|
558 |
-
visualize_results,
|
559 |
-
inputs=[results_df, text_column],
|
560 |
-
outputs=[visualization]
|
561 |
)
|
562 |
|
|
|
563 |
def create_example_data():
|
564 |
"""Create example data for demonstration"""
|
565 |
from utils import create_example_file
|
|
|
566 |
example_path = create_example_file()
|
567 |
return f"Example file created at: {example_path}"
|
568 |
|
|
|
569 |
if __name__ == "__main__":
|
570 |
# Create examples directory and sample file if it doesn't exist
|
571 |
if not os.path.exists("examples"):
|
572 |
create_example_data()
|
573 |
-
|
574 |
# Launch the Gradio app
|
575 |
demo.launch()
|
|
|
20 |
CATEGORY_SUGGESTION_PROMPT,
|
21 |
ADDITIONAL_CATEGORY_PROMPT,
|
22 |
VALIDATION_ANALYSIS_PROMPT,
|
23 |
+
CATEGORY_IMPROVEMENT_PROMPT,
|
24 |
)
|
25 |
|
26 |
# Configure logging
|
27 |
+
logging.basicConfig(
|
28 |
+
level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s"
|
29 |
+
)
|
30 |
|
31 |
# Initialize API key from environment variable
|
32 |
OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY", "")
|
|
|
40 |
except Exception as e:
|
41 |
logging.error(f"Failed to initialize OpenAI client: {str(e)}")
|
42 |
|
43 |
+
|
44 |
def update_api_key(api_key):
|
45 |
"""Update the OpenAI API key"""
|
46 |
global OPENAI_API_KEY, client
|
47 |
+
|
48 |
if not api_key:
|
49 |
return "API Key cannot be empty"
|
50 |
+
|
51 |
OPENAI_API_KEY = api_key
|
52 |
+
|
53 |
try:
|
54 |
client = OpenAI(api_key=api_key)
|
55 |
# Test the connection with a simple request
|
56 |
response = client.chat.completions.create(
|
57 |
model="gpt-3.5-turbo",
|
58 |
messages=[{"role": "user", "content": "test"}],
|
59 |
+
max_tokens=5,
|
60 |
)
|
61 |
return f"API Key updated and verified successfully"
|
62 |
except Exception as e:
|
|
|
64 |
logging.error(f"API key update failed: {error_msg}")
|
65 |
return f"Failed to update API Key: {error_msg}"
|
66 |
|
67 |
+
|
68 |
def process_file(file, text_columns, categories, classifier_type, show_explanations):
|
69 |
"""Process the uploaded file and classify text data"""
|
70 |
# Initialize result_df and validation_report
|
71 |
result_df = None
|
72 |
validation_report = None
|
73 |
+
|
74 |
try:
|
75 |
# Load data from file
|
76 |
if isinstance(file, str):
|
77 |
df = load_data(file)
|
78 |
else:
|
79 |
df = load_data(file.name)
|
80 |
+
|
81 |
if not text_columns:
|
82 |
return None, "Please select at least one text column"
|
83 |
+
|
84 |
# Check if all selected columns exist
|
85 |
missing_columns = [col for col in text_columns if col not in df.columns]
|
86 |
if missing_columns:
|
87 |
+
return (
|
88 |
+
None,
|
89 |
+
f"Columns not found in the file: {', '.join(missing_columns)}. Available columns: {', '.join(df.columns)}",
|
90 |
+
)
|
91 |
+
|
92 |
# Combine text from selected columns
|
93 |
texts = []
|
94 |
for _, row in df.iterrows():
|
95 |
combined_text = " ".join(str(row[col]) for col in text_columns)
|
96 |
texts.append(combined_text)
|
97 |
+
|
98 |
# Parse categories if provided
|
99 |
category_list = []
|
100 |
if categories:
|
101 |
category_list = [cat.strip() for cat in categories.split(",")]
|
102 |
+
|
103 |
# Select classifier based on data size and user choice
|
104 |
num_texts = len(texts)
|
105 |
+
|
106 |
# If no specific model is chosen, select the most appropriate one
|
107 |
if classifier_type == "auto":
|
108 |
if num_texts <= 500:
|
|
|
113 |
classifier_type = "hybrid"
|
114 |
else:
|
115 |
classifier_type = "tfidf"
|
116 |
+
|
117 |
# Initialize appropriate classifier
|
118 |
if classifier_type == "tfidf":
|
119 |
classifier = TFIDFClassifier()
|
120 |
results = classifier.classify(texts, category_list)
|
121 |
elif classifier_type in ["gpt35", "gpt4"]:
|
122 |
if client is None:
|
123 |
+
return (
|
124 |
+
None,
|
125 |
+
"Erreur : Le client API n'est pas initialisé. Veuillez configurer une clé API valide dans l'onglet 'Setup'.",
|
126 |
+
)
|
127 |
model = "gpt-3.5-turbo" if classifier_type == "gpt35" else "gpt-4"
|
128 |
classifier = LLMClassifier(client=client, model=model)
|
129 |
results = classifier.classify(texts, category_list)
|
130 |
else: # hybrid
|
131 |
if client is None:
|
132 |
+
return (
|
133 |
+
None,
|
134 |
+
"Erreur : Le client API n'est pas initialisé. Veuillez configurer une clé API valide dans l'onglet 'Setup'.",
|
135 |
+
)
|
136 |
# First pass with TF-IDF
|
137 |
tfidf_classifier = TFIDFClassifier()
|
138 |
tfidf_results = tfidf_classifier.classify(texts, category_list)
|
139 |
+
|
140 |
# Second pass with LLM for low confidence results
|
141 |
llm_classifier = LLMClassifier(client=client, model="gpt-3.5-turbo")
|
142 |
results = []
|
143 |
low_confidence_texts = []
|
144 |
low_confidence_indices = []
|
145 |
+
|
146 |
for i, (text, tfidf_result) in enumerate(zip(texts, tfidf_results)):
|
147 |
if tfidf_result["confidence"] < 70: # If confidence is below 70%
|
148 |
low_confidence_texts.append(text)
|
|
|
150 |
results.append(None) # Placeholder
|
151 |
else:
|
152 |
results.append(tfidf_result)
|
153 |
+
|
154 |
if low_confidence_texts:
|
155 |
+
llm_results = llm_classifier.classify(
|
156 |
+
low_confidence_texts, category_list
|
157 |
+
)
|
158 |
for idx, llm_result in zip(low_confidence_indices, llm_results):
|
159 |
results[idx] = llm_result
|
160 |
+
|
161 |
# Create results dataframe
|
162 |
result_df = df.copy()
|
163 |
result_df["Category"] = [r["category"] for r in results]
|
164 |
result_df["Confidence"] = [r["confidence"] for r in results]
|
165 |
+
|
166 |
if show_explanations:
|
167 |
result_df["Explanation"] = [r["explanation"] for r in results]
|
168 |
+
|
169 |
# Validate results using LLM
|
170 |
validation_report = validate_results(result_df, text_columns, client)
|
171 |
+
|
172 |
return result_df, validation_report
|
173 |
+
|
174 |
except Exception as e:
|
175 |
error_traceback = traceback.format_exc()
|
176 |
return None, f"Error: {str(e)}\n{error_traceback}"
|
177 |
|
178 |
+
|
179 |
def export_results(df, format_type):
|
180 |
"""Export results to a file and return the file path for download"""
|
181 |
if df is None:
|
182 |
return None
|
183 |
+
|
184 |
# Create a temporary file
|
185 |
import tempfile
|
186 |
import os
|
187 |
+
|
188 |
# Create a temporary directory if it doesn't exist
|
189 |
temp_dir = "temp_exports"
|
190 |
os.makedirs(temp_dir, exist_ok=True)
|
191 |
+
|
192 |
# Generate a unique filename
|
193 |
timestamp = time.strftime("%Y%m%d-%H%M%S")
|
194 |
filename = f"classification_results_{timestamp}"
|
195 |
+
|
196 |
if format_type == "excel":
|
197 |
file_path = os.path.join(temp_dir, f"{filename}.xlsx")
|
198 |
df.to_excel(file_path, index=False)
|
199 |
else:
|
200 |
file_path = os.path.join(temp_dir, f"{filename}.csv")
|
201 |
df.to_csv(file_path, index=False)
|
202 |
+
|
203 |
return file_path
|
204 |
|
205 |
+
|
206 |
# Create Gradio interface
|
207 |
with gr.Blocks(title="Text Classification System") as demo:
|
208 |
gr.Markdown("# Text Classification System")
|
209 |
gr.Markdown("Upload your data file (Excel/CSV) and classify text using AI")
|
210 |
+
|
211 |
with gr.Tab("Setup"):
|
212 |
api_key_input = gr.Textbox(
|
213 |
label="OpenAI API Key",
|
214 |
placeholder="Enter your API key here",
|
215 |
type="password",
|
216 |
+
value=OPENAI_API_KEY,
|
217 |
)
|
218 |
api_key_button = gr.Button("Update API Key")
|
219 |
api_key_message = gr.Textbox(label="Status", interactive=False)
|
220 |
+
|
221 |
# Display current API status
|
222 |
+
api_status = (
|
223 |
+
"API Key is set" if OPENAI_API_KEY else "No API Key found. Please set one."
|
224 |
+
)
|
225 |
gr.Markdown(f"**Current API Status**: {api_status}")
|
226 |
+
|
227 |
+
api_key_button.click(
|
228 |
+
update_api_key, inputs=[api_key_input], outputs=[api_key_message]
|
229 |
+
)
|
230 |
+
|
231 |
with gr.Tab("Classify Data"):
|
232 |
with gr.Column():
|
233 |
file_input = gr.File(label="Upload Excel/CSV File")
|
234 |
+
|
235 |
# Variable to store available columns
|
236 |
available_columns = gr.State([])
|
237 |
+
|
238 |
# Button to load file and suggest categories
|
239 |
load_categories_button = gr.Button("Load File")
|
240 |
+
|
241 |
# Display original dataframe
|
242 |
original_df = gr.Dataframe(
|
243 |
+
label="Original Data", interactive=False, visible=False
|
|
|
|
|
244 |
)
|
245 |
|
246 |
with gr.Row():
|
|
|
250 |
choices=[],
|
251 |
value=[],
|
252 |
interactive=True,
|
253 |
+
visible=False,
|
254 |
)
|
255 |
|
256 |
new_category = gr.Textbox(
|
257 |
label="Add New Category",
|
258 |
placeholder="Enter a new category name",
|
259 |
+
visible=False,
|
260 |
)
|
261 |
with gr.Row():
|
262 |
add_category_button = gr.Button("Add Category", visible=False)
|
263 |
+
suggest_category_button = gr.Button(
|
264 |
+
"Suggest Category", visible=False
|
265 |
+
)
|
266 |
|
267 |
# Original categories input (hidden)
|
268 |
+
categories = gr.Textbox(visible=False)
|
269 |
+
|
|
|
|
|
|
|
270 |
with gr.Column():
|
271 |
text_column = gr.CheckboxGroup(
|
272 |
+
label="Select Text Columns",
|
273 |
+
choices=[],
|
274 |
interactive=True,
|
275 |
+
visible=False,
|
276 |
)
|
277 |
|
278 |
classifier_type = gr.Dropdown(
|
|
|
280 |
("TF-IDF (Rapide, <1000 lignes)", "tfidf"),
|
281 |
("LLM GPT-3.5 (Fiable, <1000 lignes)", "gpt35"),
|
282 |
("LLM GPT-4 (Très fiable, <500 lignes)", "gpt4"),
|
283 |
+
("TF-IDF + LLM (Hybride, >1000 lignes)", "hybrid"),
|
284 |
],
|
285 |
label="Modèle de classification",
|
286 |
value="gpt35",
|
287 |
+
visible=False,
|
288 |
+
)
|
289 |
+
show_explanations = gr.Checkbox(
|
290 |
+
label="Show Explanations", value=True, visible=False
|
291 |
)
|
292 |
+
|
|
|
293 |
process_button = gr.Button("Process and Classify", visible=False)
|
294 |
|
295 |
results_df = gr.Dataframe(interactive=True, visible=False)
|
296 |
+
|
297 |
# Create containers for visualization and validation report
|
298 |
with gr.Row(visible=False) as results_row:
|
299 |
with gr.Column():
|
|
|
302 |
csv_download = gr.File(label="Download CSV", visible=False)
|
303 |
excel_download = gr.File(label="Download Excel", visible=False)
|
304 |
with gr.Column():
|
305 |
+
validation_output = gr.Textbox(
|
306 |
+
label="Validation Report", interactive=False
|
307 |
+
)
|
308 |
+
improve_button = gr.Button(
|
309 |
+
"Improve Classification with Report", visible=False
|
310 |
+
)
|
311 |
|
312 |
# Function to load file and suggest categories
|
313 |
def load_file_and_suggest_categories(file):
|
314 |
if not file:
|
315 |
+
return (
|
316 |
+
[],
|
317 |
+
gr.CheckboxGroup(choices=[]),
|
318 |
+
gr.CheckboxGroup(choices=[], visible=False),
|
319 |
+
gr.Textbox(visible=False),
|
320 |
+
gr.Button(visible=False),
|
321 |
+
gr.Button(visible=False),
|
322 |
+
gr.CheckboxGroup(choices=[], visible=False),
|
323 |
+
gr.Dropdown(visible=False),
|
324 |
+
gr.Checkbox(visible=False),
|
325 |
+
gr.Button(visible=False),
|
326 |
+
gr.Dataframe(visible=False),
|
327 |
+
)
|
328 |
try:
|
329 |
df = load_data(file.name)
|
330 |
columns = list(df.columns)
|
331 |
+
|
332 |
# Analyze columns to suggest text columns
|
333 |
suggested_text_columns = []
|
334 |
for col in columns:
|
335 |
# Check if column contains text data
|
336 |
+
if df[col].dtype == "object": # String type
|
337 |
# Check if column contains mostly text (not just numbers or dates)
|
338 |
sample = df[col].head(100).dropna()
|
339 |
if len(sample) > 0:
|
340 |
# Check if most values contain spaces (indicating text)
|
341 |
+
text_ratio = sum(" " in str(val) for val in sample) / len(
|
342 |
+
sample
|
343 |
+
)
|
344 |
+
if (
|
345 |
+
text_ratio > 0.3
|
346 |
+
): # If more than 30% of values contain spaces
|
347 |
suggested_text_columns.append(col)
|
348 |
+
|
349 |
# If no columns were suggested, use all object columns
|
350 |
if not suggested_text_columns:
|
351 |
+
suggested_text_columns = [
|
352 |
+
col for col in columns if df[col].dtype == "object"
|
353 |
+
]
|
354 |
+
|
355 |
# Get a sample of text for category suggestion
|
356 |
sample_texts = []
|
357 |
for col in suggested_text_columns:
|
358 |
sample_texts.extend(df[col].head(5).tolist())
|
359 |
+
|
360 |
# Use LLM to suggest categories
|
361 |
if client:
|
362 |
+
prompt = CATEGORY_SUGGESTION_PROMPT.format(
|
363 |
+
"\n---\n".join(sample_texts[:5])
|
364 |
+
)
|
365 |
try:
|
366 |
response = client.chat.completions.create(
|
367 |
model="gpt-3.5-turbo",
|
368 |
messages=[{"role": "user", "content": prompt}],
|
369 |
temperature=0,
|
370 |
+
max_tokens=100,
|
371 |
)
|
372 |
+
suggested_cats = [
|
373 |
+
cat.strip()
|
374 |
+
for cat in response.choices[0]
|
375 |
+
.message.content.strip()
|
376 |
+
.split(",")
|
377 |
+
]
|
378 |
except:
|
379 |
+
suggested_cats = [
|
380 |
+
"Positive",
|
381 |
+
"Negative",
|
382 |
+
"Neutral",
|
383 |
+
"Mixed",
|
384 |
+
"Other",
|
385 |
+
]
|
386 |
else:
|
387 |
+
suggested_cats = [
|
388 |
+
"Positive",
|
389 |
+
"Negative",
|
390 |
+
"Neutral",
|
391 |
+
"Mixed",
|
392 |
+
"Other",
|
393 |
+
]
|
394 |
+
|
395 |
return (
|
396 |
+
columns,
|
397 |
+
gr.CheckboxGroup(choices=columns, value=suggested_text_columns),
|
398 |
+
gr.CheckboxGroup(
|
399 |
+
choices=suggested_cats, value=suggested_cats, visible=True
|
400 |
+
),
|
401 |
gr.Textbox(visible=True),
|
402 |
gr.Button(visible=True),
|
403 |
gr.Button(visible=True),
|
404 |
+
gr.CheckboxGroup(
|
405 |
+
choices=columns, value=suggested_text_columns, visible=True
|
406 |
+
),
|
407 |
gr.Dropdown(visible=True),
|
408 |
gr.Checkbox(visible=True),
|
409 |
gr.Button(visible=True),
|
410 |
+
gr.Dataframe(value=df, visible=True),
|
411 |
)
|
412 |
except Exception as e:
|
413 |
+
return (
|
414 |
+
[],
|
415 |
+
gr.CheckboxGroup(choices=[]),
|
416 |
+
gr.CheckboxGroup(choices=[], visible=False),
|
417 |
+
gr.Textbox(visible=False),
|
418 |
+
gr.Button(visible=False),
|
419 |
+
gr.Button(visible=False),
|
420 |
+
gr.CheckboxGroup(choices=[], visible=False),
|
421 |
+
gr.Dropdown(visible=False),
|
422 |
+
gr.Checkbox(visible=False),
|
423 |
+
gr.Button(visible=False),
|
424 |
+
gr.Dataframe(visible=False),
|
425 |
+
)
|
426 |
+
|
427 |
# Function to add a new category
|
428 |
def add_new_category(current_categories, new_category):
|
429 |
if not new_category or new_category.strip() == "":
|
430 |
return current_categories
|
431 |
new_categories = current_categories + [new_category.strip()]
|
432 |
return gr.CheckboxGroup(choices=new_categories, value=new_categories)
|
433 |
+
|
434 |
# Function to update categories textbox
|
435 |
def update_categories_textbox(selected_categories):
|
436 |
return ", ".join(selected_categories)
|
437 |
+
|
438 |
# Function to show results after processing
|
439 |
def show_results(df, validation_report):
|
440 |
"""Show the results after processing"""
|
441 |
if df is None:
|
442 |
+
return (
|
443 |
+
gr.Row(visible=False),
|
444 |
+
gr.File(visible=False),
|
445 |
+
gr.File(visible=False),
|
446 |
+
gr.Dataframe(visible=False),
|
447 |
+
)
|
448 |
+
|
449 |
# Export to both formats
|
450 |
csv_path = export_results(df, "csv")
|
451 |
excel_path = export_results(df, "excel")
|
452 |
+
|
453 |
+
return (
|
454 |
+
gr.Row(visible=True),
|
455 |
+
gr.File(value=csv_path, visible=True),
|
456 |
+
gr.File(value=excel_path, visible=True),
|
457 |
+
gr.Dataframe(value=df, visible=True),
|
458 |
+
)
|
459 |
+
|
460 |
# Function to suggest a new category
|
461 |
def suggest_new_category(file, current_categories, text_columns):
|
462 |
if not file or not text_columns:
|
463 |
+
return gr.CheckboxGroup(
|
464 |
+
choices=current_categories, value=current_categories
|
465 |
+
)
|
466 |
+
|
467 |
try:
|
468 |
df = load_data(file.name)
|
469 |
+
|
470 |
# Get sample texts from selected columns
|
471 |
sample_texts = []
|
472 |
for col in text_columns:
|
473 |
sample_texts.extend(df[col].head(5).tolist())
|
474 |
+
|
475 |
if client:
|
476 |
prompt = ADDITIONAL_CATEGORY_PROMPT.format(
|
477 |
existing_categories=", ".join(current_categories),
|
478 |
+
sample_texts="\n---\n".join(sample_texts[:10]),
|
479 |
)
|
480 |
try:
|
481 |
response = client.chat.completions.create(
|
482 |
model="gpt-3.5-turbo",
|
483 |
messages=[{"role": "user", "content": prompt}],
|
484 |
temperature=0,
|
485 |
+
max_tokens=50,
|
486 |
)
|
487 |
new_cat = response.choices[0].message.content.strip()
|
488 |
if new_cat and new_cat not in current_categories:
|
489 |
current_categories.append(new_cat)
|
490 |
except:
|
491 |
pass
|
492 |
+
|
493 |
+
return gr.CheckboxGroup(
|
494 |
+
choices=current_categories, value=current_categories
|
495 |
+
)
|
496 |
except Exception as e:
|
497 |
+
return gr.CheckboxGroup(
|
498 |
+
choices=current_categories, value=current_categories
|
499 |
+
)
|
500 |
+
|
501 |
# Function to handle export and show download button
|
502 |
def handle_export(df, format_type):
|
503 |
if df is None:
|
504 |
return gr.File(visible=False)
|
505 |
file_path = export_results(df, format_type)
|
506 |
return gr.File(value=file_path, visible=True)
|
507 |
+
|
508 |
# Function to improve classification based on validation report
|
509 |
+
def improve_classification(
|
510 |
+
df,
|
511 |
+
validation_report,
|
512 |
+
text_columns,
|
513 |
+
categories,
|
514 |
+
classifier_type,
|
515 |
+
show_explanations,
|
516 |
+
file,
|
517 |
+
):
|
518 |
"""Improve classification based on validation report"""
|
519 |
if df is None or not validation_report:
|
520 |
+
return (
|
521 |
+
df,
|
522 |
+
validation_report,
|
523 |
+
gr.Button(visible=False),
|
524 |
+
gr.CheckboxGroup(choices=[], value=[]),
|
525 |
+
)
|
526 |
+
|
527 |
try:
|
528 |
# Extract insights from validation report
|
529 |
if client:
|
530 |
prompt = VALIDATION_ANALYSIS_PROMPT.format(
|
531 |
validation_report=validation_report,
|
532 |
+
current_categories=categories,
|
533 |
)
|
534 |
try:
|
535 |
response = client.chat.completions.create(
|
536 |
model="gpt-4",
|
537 |
messages=[{"role": "user", "content": prompt}],
|
538 |
temperature=0,
|
539 |
+
max_tokens=300,
|
540 |
+
)
|
541 |
+
improvements = json.loads(
|
542 |
+
response.choices[0].message.content.strip()
|
543 |
)
|
544 |
+
|
|
|
545 |
# Get current categories
|
546 |
+
current_categories = [
|
547 |
+
cat.strip() for cat in categories.split(",")
|
548 |
+
]
|
549 |
+
|
550 |
# If new categories are needed, suggest them based on the data
|
551 |
if improvements.get("new_categories_needed", False):
|
552 |
# Get sample texts for category suggestion
|
|
|
557 |
else:
|
558 |
temp_df = load_data(file.name)
|
559 |
sample_texts.extend(temp_df[col].head(10).tolist())
|
560 |
+
|
561 |
category_prompt = CATEGORY_IMPROVEMENT_PROMPT.format(
|
562 |
current_categories=", ".join(current_categories),
|
563 |
+
analysis=improvements.get("analysis", ""),
|
564 |
+
sample_texts="\n---\n".join(sample_texts[:10]),
|
565 |
)
|
566 |
+
|
567 |
category_response = client.chat.completions.create(
|
568 |
model="gpt-4",
|
569 |
messages=[{"role": "user", "content": category_prompt}],
|
570 |
temperature=0,
|
571 |
+
max_tokens=100,
|
572 |
)
|
573 |
+
|
574 |
+
new_categories = [
|
575 |
+
cat.strip()
|
576 |
+
for cat in category_response.choices[0]
|
577 |
+
.message.content.strip()
|
578 |
+
.split(",")
|
579 |
+
]
|
580 |
# Combine current and new categories
|
581 |
all_categories = current_categories + new_categories
|
582 |
categories = ",".join(all_categories)
|
583 |
+
|
584 |
# Process with improved parameters
|
585 |
improved_df, new_validation = process_file(
|
586 |
file,
|
587 |
text_columns,
|
588 |
categories,
|
589 |
classifier_type,
|
590 |
+
show_explanations,
|
591 |
+
)
|
592 |
+
|
593 |
+
return (
|
594 |
+
improved_df,
|
595 |
+
new_validation,
|
596 |
+
gr.Button(visible=True),
|
597 |
+
gr.CheckboxGroup(
|
598 |
+
choices=all_categories, value=all_categories
|
599 |
+
),
|
600 |
)
|
|
|
|
|
601 |
except Exception as e:
|
602 |
print(f"Error in improvement process: {str(e)}")
|
603 |
+
return (
|
604 |
+
df,
|
605 |
+
validation_report,
|
606 |
+
gr.Button(visible=True),
|
607 |
+
gr.CheckboxGroup(
|
608 |
+
choices=current_categories, value=current_categories
|
609 |
+
),
|
610 |
+
)
|
611 |
else:
|
612 |
+
return (
|
613 |
+
df,
|
614 |
+
validation_report,
|
615 |
+
gr.Button(visible=True),
|
616 |
+
gr.CheckboxGroup(
|
617 |
+
choices=current_categories, value=current_categories
|
618 |
+
),
|
619 |
+
)
|
620 |
except Exception as e:
|
621 |
print(f"Error in improvement process: {str(e)}")
|
622 |
+
return (
|
623 |
+
df,
|
624 |
+
validation_report,
|
625 |
+
gr.Button(visible=True),
|
626 |
+
gr.CheckboxGroup(
|
627 |
+
choices=current_categories, value=current_categories
|
628 |
+
),
|
629 |
+
)
|
630 |
+
|
631 |
# Connect functions
|
632 |
load_categories_button.click(
|
633 |
load_file_and_suggest_categories,
|
634 |
inputs=[file_input],
|
635 |
outputs=[
|
636 |
+
available_columns,
|
637 |
+
text_column,
|
638 |
suggested_categories,
|
639 |
new_category,
|
640 |
add_category_button,
|
|
|
643 |
classifier_type,
|
644 |
show_explanations,
|
645 |
process_button,
|
646 |
+
original_df,
|
647 |
+
],
|
648 |
)
|
649 |
+
|
650 |
add_category_button.click(
|
651 |
add_new_category,
|
652 |
inputs=[suggested_categories, new_category],
|
653 |
+
outputs=[suggested_categories],
|
654 |
)
|
655 |
+
|
656 |
suggested_categories.change(
|
657 |
update_categories_textbox,
|
658 |
inputs=[suggested_categories],
|
659 |
+
outputs=[categories],
|
660 |
)
|
661 |
+
|
662 |
suggest_category_button.click(
|
663 |
suggest_new_category,
|
664 |
inputs=[file_input, suggested_categories, text_column],
|
665 |
+
outputs=[suggested_categories],
|
666 |
)
|
667 |
+
|
668 |
process_button.click(
|
669 |
+
lambda: gr.Dataframe(visible=True), inputs=[], outputs=[results_df]
|
|
|
|
|
670 |
).then(
|
671 |
process_file,
|
672 |
+
inputs=[
|
673 |
+
file_input,
|
674 |
+
text_column,
|
675 |
+
categories,
|
676 |
+
classifier_type,
|
677 |
+
show_explanations,
|
678 |
+
],
|
679 |
+
outputs=[results_df, validation_output],
|
680 |
).then(
|
681 |
show_results,
|
682 |
inputs=[results_df, validation_output],
|
683 |
+
outputs=[results_row, csv_download, excel_download, results_df],
|
684 |
).then(
|
685 |
+
visualize_results, inputs=[results_df, text_column], outputs=[visualization]
|
|
|
|
|
686 |
).then(
|
687 |
+
lambda x: gr.Button(visible=True), inputs=[], outputs=[improve_button]
|
|
|
|
|
688 |
)
|
689 |
+
|
690 |
improve_button.click(
|
691 |
improve_classification,
|
692 |
+
inputs=[
|
693 |
+
results_df,
|
694 |
+
validation_output,
|
695 |
+
text_column,
|
696 |
+
categories,
|
697 |
+
classifier_type,
|
698 |
+
show_explanations,
|
699 |
+
file_input,
|
700 |
+
],
|
701 |
+
outputs=[
|
702 |
+
results_df,
|
703 |
+
validation_output,
|
704 |
+
improve_button,
|
705 |
+
suggested_categories,
|
706 |
+
],
|
707 |
).then(
|
708 |
show_results,
|
709 |
inputs=[results_df, validation_output],
|
710 |
+
outputs=[results_row, csv_download, excel_download, results_df],
|
711 |
).then(
|
712 |
+
visualize_results, inputs=[results_df, text_column], outputs=[visualization]
|
|
|
|
|
713 |
)
|
714 |
|
715 |
+
|
716 |
def create_example_data():
|
717 |
"""Create example data for demonstration"""
|
718 |
from utils import create_example_file
|
719 |
+
|
720 |
example_path = create_example_file()
|
721 |
return f"Example file created at: {example_path}"
|
722 |
|
723 |
+
|
724 |
if __name__ == "__main__":
|
725 |
# Create examples directory and sample file if it doesn't exist
|
726 |
if not os.path.exists("examples"):
|
727 |
create_example_data()
|
728 |
+
|
729 |
# Launch the Gradio app
|
730 |
demo.launch()
|
classifiers.py
CHANGED
@@ -9,32 +9,34 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
|
|
9 |
from typing import List, Dict, Any, Optional
|
10 |
from prompts import CATEGORY_SUGGESTION_PROMPT, TEXT_CLASSIFICATION_PROMPT
|
11 |
|
|
|
12 |
class BaseClassifier:
|
13 |
"""Base class for text classifiers"""
|
|
|
14 |
def __init__(self):
|
15 |
pass
|
16 |
-
|
17 |
def classify(self, texts, categories=None):
|
18 |
"""
|
19 |
Classify a list of texts into categories
|
20 |
-
|
21 |
Args:
|
22 |
texts (list): List of text strings to classify
|
23 |
categories (list, optional): List of category names. If None, categories will be auto-detected
|
24 |
-
|
25 |
Returns:
|
26 |
list: List of classification results with categories, confidence scores, and explanations
|
27 |
"""
|
28 |
raise NotImplementedError("Subclasses must implement this method")
|
29 |
-
|
30 |
def _generate_default_categories(self, texts, num_clusters=5):
|
31 |
"""
|
32 |
Generate default categories based on text clustering
|
33 |
-
|
34 |
Args:
|
35 |
texts (list): List of text strings
|
36 |
num_clusters (int): Number of clusters to generate
|
37 |
-
|
38 |
Returns:
|
39 |
list: List of category names
|
40 |
"""
|
@@ -45,25 +47,23 @@ class BaseClassifier:
|
|
45 |
|
46 |
class TFIDFClassifier(BaseClassifier):
|
47 |
"""Classifier using TF-IDF and clustering for fast classification"""
|
48 |
-
|
49 |
def __init__(self):
|
50 |
super().__init__()
|
51 |
self.vectorizer = TfidfVectorizer(
|
52 |
-
max_features=1000,
|
53 |
-
stop_words='english',
|
54 |
-
ngram_range=(1, 2)
|
55 |
)
|
56 |
self.model = None
|
57 |
self.feature_names = None
|
58 |
self.categories = None
|
59 |
self.centroids = None
|
60 |
-
|
61 |
def classify(self, texts, categories=None):
|
62 |
"""Classify texts using TF-IDF and clustering"""
|
63 |
# Vectorize the texts
|
64 |
X = self.vectorizer.fit_transform(texts)
|
65 |
self.feature_names = self.vectorizer.get_feature_names_out()
|
66 |
-
|
67 |
# Auto-detect categories if not provided
|
68 |
if not categories:
|
69 |
num_clusters = min(5, len(texts)) # Don't create more clusters than texts
|
@@ -71,98 +71,106 @@ class TFIDFClassifier(BaseClassifier):
|
|
71 |
else:
|
72 |
self.categories = categories
|
73 |
num_clusters = len(categories)
|
74 |
-
|
75 |
# Cluster the texts
|
76 |
self.model = KMeans(n_clusters=num_clusters, random_state=42)
|
77 |
clusters = self.model.fit_predict(X)
|
78 |
self.centroids = self.model.cluster_centers_
|
79 |
-
|
80 |
# Calculate distances to centroids for confidence
|
81 |
distances = self._calculate_distances(X)
|
82 |
-
|
83 |
# Prepare results
|
84 |
results = []
|
85 |
for i, text in enumerate(texts):
|
86 |
cluster_idx = clusters[i]
|
87 |
-
|
88 |
# Calculate confidence (inverse of distance, normalized)
|
89 |
confidence = self._calculate_confidence(distances[i])
|
90 |
-
|
91 |
# Create explanation
|
92 |
explanation = self._generate_explanation(X[i], cluster_idx)
|
93 |
-
|
94 |
-
results.append(
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
|
|
|
|
100 |
return results
|
101 |
-
|
102 |
def _calculate_distances(self, X):
|
103 |
"""Calculate distances from each point to each centroid"""
|
104 |
-
return np.sqrt(
|
105 |
-
|
|
|
|
|
|
|
|
|
106 |
def _calculate_confidence(self, distances):
|
107 |
"""Convert distances to confidence scores (0-100)"""
|
108 |
min_dist = np.min(distances)
|
109 |
max_dist = np.max(distances)
|
110 |
-
|
111 |
# Normalize and invert (smaller distance = higher confidence)
|
112 |
if max_dist == min_dist:
|
113 |
return 70 # Default mid-range confidence when all distances are equal
|
114 |
-
|
115 |
normalized_dist = (distances - min_dist) / (max_dist - min_dist)
|
116 |
min_normalized = np.min(normalized_dist)
|
117 |
-
|
118 |
# Invert and scale to 50-100 range (TF-IDF is never 100% confident)
|
119 |
confidence = 100 - (min_normalized * 50)
|
120 |
return round(confidence, 1)
|
121 |
-
|
122 |
def _generate_explanation(self, text_vector, cluster_idx):
|
123 |
"""Generate an explanation for the classification"""
|
124 |
# Get the most important features for this cluster
|
125 |
centroid = self.centroids[cluster_idx]
|
126 |
-
|
127 |
# Get indices of top features for this text
|
128 |
text_array = text_vector.toarray()[0]
|
129 |
top_indices = text_array.argsort()[-5:][::-1]
|
130 |
-
|
131 |
# Get the feature names for these indices
|
132 |
top_features = [self.feature_names[i] for i in top_indices if text_array[i] > 0]
|
133 |
-
|
134 |
if not top_features:
|
135 |
return "No significant features identified for this classification."
|
136 |
-
|
137 |
explanation = f"Classification based on key terms: {', '.join(top_features)}"
|
138 |
return explanation
|
139 |
|
140 |
|
141 |
class LLMClassifier(BaseClassifier):
|
142 |
"""Classifier using a Large Language Model for more accurate but slower classification"""
|
143 |
-
|
144 |
def __init__(self, client, model="gpt-3.5-turbo"):
|
145 |
super().__init__()
|
146 |
self.client = client
|
147 |
self.model = model
|
148 |
-
|
149 |
-
def classify(
|
|
|
|
|
150 |
"""Classify texts using an LLM with parallel processing"""
|
151 |
if not categories:
|
152 |
# First, use LLM to generate appropriate categories
|
153 |
categories = self._suggest_categories(texts)
|
154 |
-
|
155 |
# Process texts in parallel
|
156 |
with ThreadPoolExecutor(max_workers=10) as executor:
|
157 |
# Submit all tasks with their original indices
|
158 |
future_to_index = {
|
159 |
-
executor.submit(self._classify_text, text, categories): idx
|
160 |
for idx, text in enumerate(texts)
|
161 |
}
|
162 |
-
|
163 |
# Initialize results list with None values
|
164 |
results = [None] * len(texts)
|
165 |
-
|
166 |
# Collect results as they complete
|
167 |
for future in as_completed(future_to_index):
|
168 |
original_idx = future_to_index[future]
|
@@ -174,11 +182,11 @@ class LLMClassifier(BaseClassifier):
|
|
174 |
results[original_idx] = {
|
175 |
"category": categories[0],
|
176 |
"confidence": 50,
|
177 |
-
"explanation": f"Error during classification: {str(e)}"
|
178 |
}
|
179 |
-
|
180 |
return results
|
181 |
-
|
182 |
def _suggest_categories(self, texts: List[str], sample_size: int = 20) -> List[str]:
|
183 |
"""Use LLM to suggest appropriate categories for the dataset"""
|
184 |
# Take a sample of texts to avoid token limitations
|
@@ -186,54 +194,55 @@ class LLMClassifier(BaseClassifier):
|
|
186 |
sample_texts = random.sample(texts, sample_size)
|
187 |
else:
|
188 |
sample_texts = texts
|
189 |
-
|
190 |
prompt = CATEGORY_SUGGESTION_PROMPT.format("\n---\n".join(sample_texts))
|
191 |
-
|
192 |
try:
|
193 |
response = self.client.chat.completions.create(
|
194 |
model=self.model,
|
195 |
messages=[{"role": "user", "content": prompt}],
|
196 |
temperature=0.2,
|
197 |
-
max_tokens=100
|
198 |
)
|
199 |
-
|
200 |
# Parse response to get categories
|
201 |
categories_text = response.choices[0].message.content.strip()
|
202 |
categories = [cat.strip() for cat in categories_text.split(",")]
|
203 |
-
|
204 |
return categories
|
205 |
except Exception as e:
|
206 |
# Fallback to default categories on error
|
207 |
print(f"Error suggesting categories: {str(e)}")
|
208 |
return self._generate_default_categories(texts)
|
209 |
-
|
210 |
def _classify_text(self, text: str, categories: List[str]) -> Dict[str, Any]:
|
211 |
"""Use LLM to classify a single text"""
|
212 |
prompt = TEXT_CLASSIFICATION_PROMPT.format(
|
213 |
-
categories=", ".join(categories),
|
214 |
-
text=text
|
215 |
)
|
216 |
-
|
217 |
try:
|
218 |
response = self.client.chat.completions.create(
|
219 |
model=self.model,
|
220 |
messages=[{"role": "user", "content": prompt}],
|
221 |
temperature=0,
|
222 |
-
max_tokens=200
|
223 |
)
|
224 |
-
|
225 |
# Parse JSON response
|
226 |
response_text = response.choices[0].message.content.strip()
|
227 |
-
|
228 |
result = json.loads(response_text)
|
229 |
# Ensure all required fields are present
|
230 |
if not all(k in result for k in ["category", "confidence", "explanation"]):
|
231 |
raise ValueError("Missing required fields in LLM response")
|
232 |
-
|
233 |
# Validate category is in the list
|
234 |
if result["category"] not in categories:
|
235 |
-
result["category"] = categories[
|
236 |
-
|
|
|
|
|
237 |
# Validate confidence is a number between 0 and 100
|
238 |
try:
|
239 |
result["confidence"] = float(result["confidence"])
|
@@ -241,7 +250,7 @@ class LLMClassifier(BaseClassifier):
|
|
241 |
result["confidence"] = 50
|
242 |
except:
|
243 |
result["confidence"] = 50
|
244 |
-
|
245 |
return result
|
246 |
except json.JSONDecodeError:
|
247 |
# Fall back to simple parsing if JSON fails
|
@@ -250,12 +259,9 @@ class LLMClassifier(BaseClassifier):
|
|
250 |
if cat.lower() in response_text.lower():
|
251 |
category = cat
|
252 |
break
|
253 |
-
|
254 |
return {
|
255 |
"category": category,
|
256 |
"confidence": 50,
|
257 |
-
"explanation": f"Classification based on language model analysis. (Note: Structured response parsing failed)"
|
258 |
}
|
259 |
-
|
260 |
-
|
261 |
-
|
|
|
9 |
from typing import List, Dict, Any, Optional
|
10 |
from prompts import CATEGORY_SUGGESTION_PROMPT, TEXT_CLASSIFICATION_PROMPT
|
11 |
|
12 |
+
|
13 |
class BaseClassifier:
|
14 |
"""Base class for text classifiers"""
|
15 |
+
|
16 |
def __init__(self):
|
17 |
pass
|
18 |
+
|
19 |
def classify(self, texts, categories=None):
|
20 |
"""
|
21 |
Classify a list of texts into categories
|
22 |
+
|
23 |
Args:
|
24 |
texts (list): List of text strings to classify
|
25 |
categories (list, optional): List of category names. If None, categories will be auto-detected
|
26 |
+
|
27 |
Returns:
|
28 |
list: List of classification results with categories, confidence scores, and explanations
|
29 |
"""
|
30 |
raise NotImplementedError("Subclasses must implement this method")
|
31 |
+
|
32 |
def _generate_default_categories(self, texts, num_clusters=5):
|
33 |
"""
|
34 |
Generate default categories based on text clustering
|
35 |
+
|
36 |
Args:
|
37 |
texts (list): List of text strings
|
38 |
num_clusters (int): Number of clusters to generate
|
39 |
+
|
40 |
Returns:
|
41 |
list: List of category names
|
42 |
"""
|
|
|
47 |
|
48 |
class TFIDFClassifier(BaseClassifier):
|
49 |
"""Classifier using TF-IDF and clustering for fast classification"""
|
50 |
+
|
51 |
def __init__(self):
|
52 |
super().__init__()
|
53 |
self.vectorizer = TfidfVectorizer(
|
54 |
+
max_features=1000, stop_words="english", ngram_range=(1, 2)
|
|
|
|
|
55 |
)
|
56 |
self.model = None
|
57 |
self.feature_names = None
|
58 |
self.categories = None
|
59 |
self.centroids = None
|
60 |
+
|
61 |
def classify(self, texts, categories=None):
|
62 |
"""Classify texts using TF-IDF and clustering"""
|
63 |
# Vectorize the texts
|
64 |
X = self.vectorizer.fit_transform(texts)
|
65 |
self.feature_names = self.vectorizer.get_feature_names_out()
|
66 |
+
|
67 |
# Auto-detect categories if not provided
|
68 |
if not categories:
|
69 |
num_clusters = min(5, len(texts)) # Don't create more clusters than texts
|
|
|
71 |
else:
|
72 |
self.categories = categories
|
73 |
num_clusters = len(categories)
|
74 |
+
|
75 |
# Cluster the texts
|
76 |
self.model = KMeans(n_clusters=num_clusters, random_state=42)
|
77 |
clusters = self.model.fit_predict(X)
|
78 |
self.centroids = self.model.cluster_centers_
|
79 |
+
|
80 |
# Calculate distances to centroids for confidence
|
81 |
distances = self._calculate_distances(X)
|
82 |
+
|
83 |
# Prepare results
|
84 |
results = []
|
85 |
for i, text in enumerate(texts):
|
86 |
cluster_idx = clusters[i]
|
87 |
+
|
88 |
# Calculate confidence (inverse of distance, normalized)
|
89 |
confidence = self._calculate_confidence(distances[i])
|
90 |
+
|
91 |
# Create explanation
|
92 |
explanation = self._generate_explanation(X[i], cluster_idx)
|
93 |
+
|
94 |
+
results.append(
|
95 |
+
{
|
96 |
+
"category": self.categories[cluster_idx],
|
97 |
+
"confidence": confidence,
|
98 |
+
"explanation": explanation,
|
99 |
+
}
|
100 |
+
)
|
101 |
+
|
102 |
return results
|
103 |
+
|
104 |
def _calculate_distances(self, X):
|
105 |
"""Calculate distances from each point to each centroid"""
|
106 |
+
return np.sqrt(
|
107 |
+
(
|
108 |
+
(X.toarray()[:, np.newaxis, :] - self.centroids[np.newaxis, :, :]) ** 2
|
109 |
+
).sum(axis=2)
|
110 |
+
)
|
111 |
+
|
112 |
def _calculate_confidence(self, distances):
|
113 |
"""Convert distances to confidence scores (0-100)"""
|
114 |
min_dist = np.min(distances)
|
115 |
max_dist = np.max(distances)
|
116 |
+
|
117 |
# Normalize and invert (smaller distance = higher confidence)
|
118 |
if max_dist == min_dist:
|
119 |
return 70 # Default mid-range confidence when all distances are equal
|
120 |
+
|
121 |
normalized_dist = (distances - min_dist) / (max_dist - min_dist)
|
122 |
min_normalized = np.min(normalized_dist)
|
123 |
+
|
124 |
# Invert and scale to 50-100 range (TF-IDF is never 100% confident)
|
125 |
confidence = 100 - (min_normalized * 50)
|
126 |
return round(confidence, 1)
|
127 |
+
|
128 |
def _generate_explanation(self, text_vector, cluster_idx):
|
129 |
"""Generate an explanation for the classification"""
|
130 |
# Get the most important features for this cluster
|
131 |
centroid = self.centroids[cluster_idx]
|
132 |
+
|
133 |
# Get indices of top features for this text
|
134 |
text_array = text_vector.toarray()[0]
|
135 |
top_indices = text_array.argsort()[-5:][::-1]
|
136 |
+
|
137 |
# Get the feature names for these indices
|
138 |
top_features = [self.feature_names[i] for i in top_indices if text_array[i] > 0]
|
139 |
+
|
140 |
if not top_features:
|
141 |
return "No significant features identified for this classification."
|
142 |
+
|
143 |
explanation = f"Classification based on key terms: {', '.join(top_features)}"
|
144 |
return explanation
|
145 |
|
146 |
|
147 |
class LLMClassifier(BaseClassifier):
|
148 |
"""Classifier using a Large Language Model for more accurate but slower classification"""
|
149 |
+
|
150 |
def __init__(self, client, model="gpt-3.5-turbo"):
|
151 |
super().__init__()
|
152 |
self.client = client
|
153 |
self.model = model
|
154 |
+
|
155 |
+
def classify(
|
156 |
+
self, texts: List[str], categories: Optional[List[str]] = None
|
157 |
+
) -> List[Dict[str, Any]]:
|
158 |
"""Classify texts using an LLM with parallel processing"""
|
159 |
if not categories:
|
160 |
# First, use LLM to generate appropriate categories
|
161 |
categories = self._suggest_categories(texts)
|
162 |
+
|
163 |
# Process texts in parallel
|
164 |
with ThreadPoolExecutor(max_workers=10) as executor:
|
165 |
# Submit all tasks with their original indices
|
166 |
future_to_index = {
|
167 |
+
executor.submit(self._classify_text, text, categories): idx
|
168 |
for idx, text in enumerate(texts)
|
169 |
}
|
170 |
+
|
171 |
# Initialize results list with None values
|
172 |
results = [None] * len(texts)
|
173 |
+
|
174 |
# Collect results as they complete
|
175 |
for future in as_completed(future_to_index):
|
176 |
original_idx = future_to_index[future]
|
|
|
182 |
results[original_idx] = {
|
183 |
"category": categories[0],
|
184 |
"confidence": 50,
|
185 |
+
"explanation": f"Error during classification: {str(e)}",
|
186 |
}
|
187 |
+
|
188 |
return results
|
189 |
+
|
190 |
def _suggest_categories(self, texts: List[str], sample_size: int = 20) -> List[str]:
|
191 |
"""Use LLM to suggest appropriate categories for the dataset"""
|
192 |
# Take a sample of texts to avoid token limitations
|
|
|
194 |
sample_texts = random.sample(texts, sample_size)
|
195 |
else:
|
196 |
sample_texts = texts
|
197 |
+
|
198 |
prompt = CATEGORY_SUGGESTION_PROMPT.format("\n---\n".join(sample_texts))
|
199 |
+
|
200 |
try:
|
201 |
response = self.client.chat.completions.create(
|
202 |
model=self.model,
|
203 |
messages=[{"role": "user", "content": prompt}],
|
204 |
temperature=0.2,
|
205 |
+
max_tokens=100,
|
206 |
)
|
207 |
+
|
208 |
# Parse response to get categories
|
209 |
categories_text = response.choices[0].message.content.strip()
|
210 |
categories = [cat.strip() for cat in categories_text.split(",")]
|
211 |
+
|
212 |
return categories
|
213 |
except Exception as e:
|
214 |
# Fallback to default categories on error
|
215 |
print(f"Error suggesting categories: {str(e)}")
|
216 |
return self._generate_default_categories(texts)
|
217 |
+
|
218 |
def _classify_text(self, text: str, categories: List[str]) -> Dict[str, Any]:
|
219 |
"""Use LLM to classify a single text"""
|
220 |
prompt = TEXT_CLASSIFICATION_PROMPT.format(
|
221 |
+
categories=", ".join(categories), text=text
|
|
|
222 |
)
|
223 |
+
|
224 |
try:
|
225 |
response = self.client.chat.completions.create(
|
226 |
model=self.model,
|
227 |
messages=[{"role": "user", "content": prompt}],
|
228 |
temperature=0,
|
229 |
+
max_tokens=200,
|
230 |
)
|
231 |
+
|
232 |
# Parse JSON response
|
233 |
response_text = response.choices[0].message.content.strip()
|
234 |
+
|
235 |
result = json.loads(response_text)
|
236 |
# Ensure all required fields are present
|
237 |
if not all(k in result for k in ["category", "confidence", "explanation"]):
|
238 |
raise ValueError("Missing required fields in LLM response")
|
239 |
+
|
240 |
# Validate category is in the list
|
241 |
if result["category"] not in categories:
|
242 |
+
result["category"] = categories[
|
243 |
+
0
|
244 |
+
] # Default to first category if invalid
|
245 |
+
|
246 |
# Validate confidence is a number between 0 and 100
|
247 |
try:
|
248 |
result["confidence"] = float(result["confidence"])
|
|
|
250 |
result["confidence"] = 50
|
251 |
except:
|
252 |
result["confidence"] = 50
|
253 |
+
|
254 |
return result
|
255 |
except json.JSONDecodeError:
|
256 |
# Fall back to simple parsing if JSON fails
|
|
|
259 |
if cat.lower() in response_text.lower():
|
260 |
category = cat
|
261 |
break
|
262 |
+
|
263 |
return {
|
264 |
"category": category,
|
265 |
"confidence": 50,
|
266 |
+
"explanation": f"Classification based on language model analysis. (Note: Structured response parsing failed)",
|
267 |
}
|
|
|
|
|
|
prompts.py
CHANGED
@@ -60,4 +60,4 @@ Example texts:
|
|
60 |
{}
|
61 |
|
62 |
Return your answer as a comma-separated list of new category names only.
|
63 |
-
"""
|
|
|
60 |
{}
|
61 |
|
62 |
Return your answer as a comma-separated list of new category names only.
|
63 |
+
"""
|
utils.py
CHANGED
@@ -6,61 +6,66 @@ from sklearn.decomposition import PCA
|
|
6 |
from sklearn.feature_extraction.text import TfidfVectorizer
|
7 |
import tempfile
|
8 |
|
|
|
9 |
def load_data(file_path):
|
10 |
"""
|
11 |
Load data from an Excel or CSV file
|
12 |
-
|
13 |
Args:
|
14 |
file_path (str): Path to the file
|
15 |
-
|
16 |
Returns:
|
17 |
pd.DataFrame: Loaded data
|
18 |
"""
|
19 |
file_ext = os.path.splitext(file_path)[1].lower()
|
20 |
-
|
21 |
-
if file_ext ==
|
22 |
return pd.read_excel(file_path)
|
23 |
-
elif file_ext ==
|
24 |
return pd.read_csv(file_path)
|
25 |
else:
|
26 |
-
raise ValueError(
|
|
|
|
|
|
|
27 |
|
28 |
def export_data(df, file_name, format_type="excel"):
|
29 |
"""
|
30 |
Export dataframe to file
|
31 |
-
|
32 |
Args:
|
33 |
df (pd.DataFrame): Dataframe to export
|
34 |
file_name (str): Name of the output file
|
35 |
format_type (str): "excel" or "csv"
|
36 |
-
|
37 |
Returns:
|
38 |
str: Path to the exported file
|
39 |
"""
|
40 |
# Create export directory if it doesn't exist
|
41 |
export_dir = "exports"
|
42 |
os.makedirs(export_dir, exist_ok=True)
|
43 |
-
|
44 |
# Full path for the export file
|
45 |
export_path = os.path.join(export_dir, file_name)
|
46 |
-
|
47 |
# Export based on format type
|
48 |
if format_type == "excel":
|
49 |
df.to_excel(export_path, index=False)
|
50 |
else:
|
51 |
df.to_csv(export_path, index=False)
|
52 |
-
|
53 |
return export_path
|
54 |
|
|
|
55 |
def visualize_results(df, text_column, category_column="Category"):
|
56 |
"""
|
57 |
Create visualization of classification results
|
58 |
-
|
59 |
Args:
|
60 |
df (pd.DataFrame): Dataframe with classification results
|
61 |
text_column (str): Name of the column containing text data
|
62 |
category_column (str): Name of the column containing categories
|
63 |
-
|
64 |
Returns:
|
65 |
matplotlib.figure.Figure: Visualization figure
|
66 |
"""
|
@@ -68,52 +73,58 @@ def visualize_results(df, text_column, category_column="Category"):
|
|
68 |
if category_column not in df.columns:
|
69 |
# Create a simple figure with a message
|
70 |
fig, ax = plt.subplots(figsize=(10, 6))
|
71 |
-
ax.text(
|
72 |
-
|
73 |
-
|
|
|
74 |
plt.tight_layout()
|
75 |
return fig
|
76 |
-
|
77 |
# Get categories and their counts
|
78 |
category_counts = df[category_column].value_counts()
|
79 |
-
|
80 |
# Create a new figure
|
81 |
fig, ax = plt.subplots(figsize=(10, 6))
|
82 |
-
|
83 |
# Create the histogram
|
84 |
bars = ax.bar(category_counts.index, category_counts.values)
|
85 |
-
|
86 |
# Add value labels on top of each bar
|
87 |
for bar in bars:
|
88 |
height = bar.get_height()
|
89 |
-
ax.text(
|
90 |
-
|
91 |
-
|
92 |
-
|
|
|
|
|
|
|
|
|
93 |
# Customize the plot
|
94 |
-
ax.set_xlabel(
|
95 |
-
ax.set_ylabel(
|
96 |
-
ax.set_title(
|
97 |
-
|
98 |
# Rotate x-axis labels if they're too long
|
99 |
-
plt.xticks(rotation=45, ha=
|
100 |
-
|
101 |
# Add grid
|
102 |
-
ax.grid(True, linestyle=
|
103 |
-
|
104 |
plt.tight_layout()
|
105 |
-
|
106 |
return fig
|
107 |
|
|
|
108 |
def validate_results(df, text_columns, client):
|
109 |
"""
|
110 |
Use LLM to validate the classification results
|
111 |
-
|
112 |
Args:
|
113 |
df (pd.DataFrame): Dataframe with classification results
|
114 |
text_columns (list): List of column names containing text data
|
115 |
client: LiteLLM client
|
116 |
-
|
117 |
Returns:
|
118 |
str: Validation report
|
119 |
"""
|
@@ -121,7 +132,7 @@ def validate_results(df, text_columns, client):
|
|
121 |
# Sample a few rows for validation
|
122 |
sample_size = min(5, len(df))
|
123 |
sample_df = df.sample(n=sample_size, random_state=42)
|
124 |
-
|
125 |
# Build validation prompt
|
126 |
validation_prompts = []
|
127 |
for _, row in sample_df.iterrows():
|
@@ -129,11 +140,11 @@ def validate_results(df, text_columns, client):
|
|
129 |
text = " ".join(str(row[col]) for col in text_columns)
|
130 |
assigned_category = row["Category"]
|
131 |
confidence = row["Confidence"]
|
132 |
-
|
133 |
validation_prompts.append(
|
134 |
f"Text: {text}\nAssigned Category: {assigned_category}\nConfidence: {confidence}\n"
|
135 |
)
|
136 |
-
|
137 |
prompt = """
|
138 |
As a validation expert, review the following text classifications and provide feedback.
|
139 |
For each text, assess whether the assigned category seems appropriate:
|
@@ -146,19 +157,21 @@ def validate_results(df, text_columns, client):
|
|
146 |
3. Suggestions for improvement
|
147 |
|
148 |
Keep your response under 300 words.
|
149 |
-
""".format(
|
150 |
-
|
|
|
|
|
151 |
# Call LLM API
|
152 |
response = client.chat.completions.create(
|
153 |
model="gpt-3.5-turbo",
|
154 |
messages=[{"role": "user", "content": prompt}],
|
155 |
temperature=0.3,
|
156 |
-
max_tokens=400
|
157 |
)
|
158 |
-
|
159 |
validation_report = response.choices[0].message.content.strip()
|
160 |
return validation_report
|
161 |
-
|
162 |
except Exception as e:
|
163 |
return f"Validation failed: {str(e)}"
|
164 |
|
@@ -166,7 +179,7 @@ def validate_results(df, text_columns, client):
|
|
166 |
def create_example_file():
|
167 |
"""
|
168 |
Create an example CSV file for testing
|
169 |
-
|
170 |
Returns:
|
171 |
str: Path to the created file
|
172 |
"""
|
@@ -182,17 +195,17 @@ def create_example_file():
|
|
182 |
"It's okay, nothing special but gets the job done.",
|
183 |
"I'm extremely disappointed with the quality of this product.",
|
184 |
"This is the best purchase I've made all year!",
|
185 |
-
"It's reasonably priced and works as expected."
|
186 |
]
|
187 |
}
|
188 |
-
|
189 |
# Create dataframe
|
190 |
df = pd.DataFrame(data)
|
191 |
-
|
192 |
# Save to a CSV file
|
193 |
example_dir = "examples"
|
194 |
os.makedirs(example_dir, exist_ok=True)
|
195 |
file_path = os.path.join(example_dir, "sample_reviews.csv")
|
196 |
df.to_csv(file_path, index=False)
|
197 |
-
|
198 |
return file_path
|
|
|
6 |
from sklearn.feature_extraction.text import TfidfVectorizer
|
7 |
import tempfile
|
8 |
|
9 |
+
|
10 |
def load_data(file_path):
|
11 |
"""
|
12 |
Load data from an Excel or CSV file
|
13 |
+
|
14 |
Args:
|
15 |
file_path (str): Path to the file
|
16 |
+
|
17 |
Returns:
|
18 |
pd.DataFrame: Loaded data
|
19 |
"""
|
20 |
file_ext = os.path.splitext(file_path)[1].lower()
|
21 |
+
|
22 |
+
if file_ext == ".xlsx" or file_ext == ".xls":
|
23 |
return pd.read_excel(file_path)
|
24 |
+
elif file_ext == ".csv":
|
25 |
return pd.read_csv(file_path)
|
26 |
else:
|
27 |
+
raise ValueError(
|
28 |
+
f"Unsupported file format: {file_ext}. Please upload an Excel or CSV file."
|
29 |
+
)
|
30 |
+
|
31 |
|
32 |
def export_data(df, file_name, format_type="excel"):
|
33 |
"""
|
34 |
Export dataframe to file
|
35 |
+
|
36 |
Args:
|
37 |
df (pd.DataFrame): Dataframe to export
|
38 |
file_name (str): Name of the output file
|
39 |
format_type (str): "excel" or "csv"
|
40 |
+
|
41 |
Returns:
|
42 |
str: Path to the exported file
|
43 |
"""
|
44 |
# Create export directory if it doesn't exist
|
45 |
export_dir = "exports"
|
46 |
os.makedirs(export_dir, exist_ok=True)
|
47 |
+
|
48 |
# Full path for the export file
|
49 |
export_path = os.path.join(export_dir, file_name)
|
50 |
+
|
51 |
# Export based on format type
|
52 |
if format_type == "excel":
|
53 |
df.to_excel(export_path, index=False)
|
54 |
else:
|
55 |
df.to_csv(export_path, index=False)
|
56 |
+
|
57 |
return export_path
|
58 |
|
59 |
+
|
60 |
def visualize_results(df, text_column, category_column="Category"):
|
61 |
"""
|
62 |
Create visualization of classification results
|
63 |
+
|
64 |
Args:
|
65 |
df (pd.DataFrame): Dataframe with classification results
|
66 |
text_column (str): Name of the column containing text data
|
67 |
category_column (str): Name of the column containing categories
|
68 |
+
|
69 |
Returns:
|
70 |
matplotlib.figure.Figure: Visualization figure
|
71 |
"""
|
|
|
73 |
if category_column not in df.columns:
|
74 |
# Create a simple figure with a message
|
75 |
fig, ax = plt.subplots(figsize=(10, 6))
|
76 |
+
ax.text(
|
77 |
+
0.5, 0.5, "No categories to display", ha="center", va="center", fontsize=12
|
78 |
+
)
|
79 |
+
ax.set_title("No Classification Results Available")
|
80 |
plt.tight_layout()
|
81 |
return fig
|
82 |
+
|
83 |
# Get categories and their counts
|
84 |
category_counts = df[category_column].value_counts()
|
85 |
+
|
86 |
# Create a new figure
|
87 |
fig, ax = plt.subplots(figsize=(10, 6))
|
88 |
+
|
89 |
# Create the histogram
|
90 |
bars = ax.bar(category_counts.index, category_counts.values)
|
91 |
+
|
92 |
# Add value labels on top of each bar
|
93 |
for bar in bars:
|
94 |
height = bar.get_height()
|
95 |
+
ax.text(
|
96 |
+
bar.get_x() + bar.get_width() / 2.0,
|
97 |
+
height,
|
98 |
+
f"{int(height)}",
|
99 |
+
ha="center",
|
100 |
+
va="bottom",
|
101 |
+
)
|
102 |
+
|
103 |
# Customize the plot
|
104 |
+
ax.set_xlabel("Categories")
|
105 |
+
ax.set_ylabel("Number of Texts")
|
106 |
+
ax.set_title("Distribution of Classified Texts")
|
107 |
+
|
108 |
# Rotate x-axis labels if they're too long
|
109 |
+
plt.xticks(rotation=45, ha="right")
|
110 |
+
|
111 |
# Add grid
|
112 |
+
ax.grid(True, linestyle="--", alpha=0.7)
|
113 |
+
|
114 |
plt.tight_layout()
|
115 |
+
|
116 |
return fig
|
117 |
|
118 |
+
|
119 |
def validate_results(df, text_columns, client):
|
120 |
"""
|
121 |
Use LLM to validate the classification results
|
122 |
+
|
123 |
Args:
|
124 |
df (pd.DataFrame): Dataframe with classification results
|
125 |
text_columns (list): List of column names containing text data
|
126 |
client: LiteLLM client
|
127 |
+
|
128 |
Returns:
|
129 |
str: Validation report
|
130 |
"""
|
|
|
132 |
# Sample a few rows for validation
|
133 |
sample_size = min(5, len(df))
|
134 |
sample_df = df.sample(n=sample_size, random_state=42)
|
135 |
+
|
136 |
# Build validation prompt
|
137 |
validation_prompts = []
|
138 |
for _, row in sample_df.iterrows():
|
|
|
140 |
text = " ".join(str(row[col]) for col in text_columns)
|
141 |
assigned_category = row["Category"]
|
142 |
confidence = row["Confidence"]
|
143 |
+
|
144 |
validation_prompts.append(
|
145 |
f"Text: {text}\nAssigned Category: {assigned_category}\nConfidence: {confidence}\n"
|
146 |
)
|
147 |
+
|
148 |
prompt = """
|
149 |
As a validation expert, review the following text classifications and provide feedback.
|
150 |
For each text, assess whether the assigned category seems appropriate:
|
|
|
157 |
3. Suggestions for improvement
|
158 |
|
159 |
Keep your response under 300 words.
|
160 |
+
""".format(
|
161 |
+
"\n---\n".join(validation_prompts)
|
162 |
+
)
|
163 |
+
|
164 |
# Call LLM API
|
165 |
response = client.chat.completions.create(
|
166 |
model="gpt-3.5-turbo",
|
167 |
messages=[{"role": "user", "content": prompt}],
|
168 |
temperature=0.3,
|
169 |
+
max_tokens=400,
|
170 |
)
|
171 |
+
|
172 |
validation_report = response.choices[0].message.content.strip()
|
173 |
return validation_report
|
174 |
+
|
175 |
except Exception as e:
|
176 |
return f"Validation failed: {str(e)}"
|
177 |
|
|
|
179 |
def create_example_file():
|
180 |
"""
|
181 |
Create an example CSV file for testing
|
182 |
+
|
183 |
Returns:
|
184 |
str: Path to the created file
|
185 |
"""
|
|
|
195 |
"It's okay, nothing special but gets the job done.",
|
196 |
"I'm extremely disappointed with the quality of this product.",
|
197 |
"This is the best purchase I've made all year!",
|
198 |
+
"It's reasonably priced and works as expected.",
|
199 |
]
|
200 |
}
|
201 |
+
|
202 |
# Create dataframe
|
203 |
df = pd.DataFrame(data)
|
204 |
+
|
205 |
# Save to a CSV file
|
206 |
example_dir = "examples"
|
207 |
os.makedirs(example_dir, exist_ok=True)
|
208 |
file_path = os.path.join(example_dir, "sample_reviews.csv")
|
209 |
df.to_csv(file_path, index=False)
|
210 |
+
|
211 |
return file_path
|