write the app.py and upload data

.gitattributes

@@ -32,3 +32,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.json filter=lfs diff=lfs merge=lfs -text
+dialogues_set/* filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,192 @@
+import openai
+import tiktoken
+
+openai.api_key = os.getenv('API_KEY')
+
+import collections
+import datetime
+import time
+import json
+import os
+
+timezone = datetime.tz.gettz('Asia/Shanghai')
+timestamp2string = lambda timestamp: datetime.datetime.fromtimestamp(timestamp).replace(tzinfo=timezone).strftime('%Y-%m-%d %H:%M:%S')
+
+def num_tokens_from_messages(messages, model="gpt-3.5-turbo"):
+    """Returns the number of tokens used by a list of messages."""
+    try:
+        encoding = tiktoken.encoding_for_model(model)
+    except KeyError:
+        encoding = tiktoken.get_encoding("cl100k_base")
+    if model == "gpt-3.5-turbo":  # note: future models may deviate from this
+        num_tokens = 0
+        for message in messages:
+            num_tokens += 4  # every message follows <im_start>{role/name}\n{content}<im_end>\n
+            for key, value in message.items():
+                num_tokens += len(encoding.encode(value))
+                if key == "name":  # if there's a name, the role is omitted
+                    num_tokens += -1  # role is always required and always 1 token
+        num_tokens += 2  # every reply is primed with <im_start>assistant
+        return num_tokens
+    else:
+        raise NotImplementedError(f"""num_tokens_from_messages() is not presently implemented for model {model}.
+See https://github.com/openai/openai-python/blob/main/chatml.md for information on how messages are converted to tokens.""")
+
+qas = []
+qs = []
+start_time = None
+
+def read_qs():
+    directory = "./dialogues_set"
+    filenames = [
+        'dialogues_film.json', 
+        'dialogues_jindong.json', 
+        'dialogues_music.json', 
+        'dialogues_natural.json', 
+        'dialogues_taobao.json', 
+        'dialogues_travel_kd.json'
+    ]
+    for filename in filenames:
+        with open(f"{directory}/{filename}", "r", encoding="utf-8") as f:
+            for idx,line in enumerate(f):
+                idx2query = json.loads(line)
+                query = idx2query[str(idx)]
+                qs.append(query)
+    print(f"read {len(qs)} queries from files")
+    
+    if os.path.exists("qas.json"):
+        with open("qas.json", "r", encoding="utf-8") as f:
+            qas = json.loads(f.read())
+        print(f"read {len(qas)} query-responses from qas.json")
+        
+        existed_qs = collections.Counter([qa["q"] for qa in qas])
+        remained_qs = []
+        for q in qs:
+            if existed_qs[q]>0:
+                existed_qs[q] -= 1
+            else:
+                remained_qs.append(q)
+        print(f"filter out {len(qs)-len(remained_qs)} with reference to qas.json")
+        qs = remained_qs
+    
+    return qs
+
+
+def ask(query, max_attempt_times=3):
+    answer = None
+    attempt_times = 0
+    while answer is None and attempt_times<max_attempt_times:
+        attempt_times += 1
+        try:
+            response = openai.ChatCompletion.create(
+                model="gpt-3.5-turbo",
+                messages=[
+                    {"role": "user", "content": query}
+                ]
+            )["choices"][0]["message"]["content"]
+        except Exception as e:
+            print(e)
+            print(f"retry in {attempt_times*10} seconds...")
+            time.sleep(attempt_times*10)
+    return answer
+
+
+def askingChatGPT(min_interval_seconds=10):
+    qs = read_qs()
+    
+    start_time = time.time()
+    for i, q in enumerate(qs):
+        ask_start_time = time.time()
+        
+        a = ask(q)
+        qas.append({"q":q, "a":a})
+        
+        ask_end_time = time.time()
+        elapsed_time = ask_end_time - ask_start_time
+        delayTime = min_interval_seconds - elapsed_time
+        if delayTime>0:
+            time.sleep(delayTime)
+        
+        print(f"{timestamp2string(time.time())}:  iterations:  {i} / {len(qs)} | elapsed time of this query (s):  {elapsed_time:.2f}")
+    
+    return
+
+
+thread = threading.Thread(target=my_function)
+thread.daemon = True
+thread.start()
+
+
+import gradio as gr
+
+
+def showcase(api_key):
+    if not api_key==openai.api_key:
+        chatbot_ret = [(f"Your entered api_key:<br>{api_key}<br>is incorrect.", f"So i cannot provide you any information in this private space.")]
+    else:
+        recent_qas = qas[:10]
+        chatbot_ret = [(f"Your entered api_key is correct.", f"The latest {len(recent_qas)} query-responses are displayed below.")]
+        for qa in recent_qas:
+            chatbot_ret += [(qa["q"],qa["a"])]
+    return chatbot_ret
+
+
+def download(api_key):
+    if not api_key==openai.api_key:
+        chatbot_ret = [(f"Your entered api_key:<br>{api_key}<br>is incorrect.", f"So i cannot provide you any information in this private space.")]
+        file_ret = gr.File.update(value=None, visible=False)
+    else:
+        chatbot_ret = [(f"Your entered api_key is correct.", f"The file containing all processed query-responses ({len(qas)} in total) can be downloaded below.")]
+        filename = f"qas{len(qas)}.json"
+        with open(filename, "w", encoding="utf-8") as f:
+            f.write(json.dumps(qas, ensure_ascii=False, indent=2))
+        file_ret = gr.File.update(value=filename, visible=True)
+    return chatbot_ret, file_ret
+
+
+def display(api_key):
+    if not api_key==openai.api_key:
+        chatbot_ret = [(f"Your entered api_key:<br>{api_key}<br>is incorrect.", f"So i cannot provide you any information in this private space.")]
+    elif len(qas)<1:
+        chatbot_ret = [(f"Your entered api_key is correct.", f"But the progress has just started for a while and has no useful progress information to provide.")]
+    else:
+        time_takes = time.time() - start_time
+        time_remains = time_takes * (len(qs)-len(qas)) / len(qas)
+        end_time = start_time + time_remains
+        
+        messages = []
+        for qa in qas:
+            messages.append({"role":"user", "content":qas["q"]})
+            messages.append({"role":"assistant", "content":qas["a"]})
+        num_tokens_processed = num_tokens_from_messages(messages)
+        num_tokens_total = num_tokens_processed * len(qs) / len(qas)
+        dollars_tokens_processed = 0.002 * int(num_tokens_processed/1000)
+        dollars_tokens_total = 0.002 * int(num_tokens_total/1000)
+        
+        chatbot_ret = [(f"Your entered api_key is correct.", f"The information of progress is displayed below.")]
+        chatbot_ret += [(f"The number of processed / total queries:", f"{len(qas)} / {len(qs)}")]
+        chatbot_ret += [(f"The hours already takes / est. remains:", f"{time_takes/3600:.2f} / {time_remains/3600:.2f}")]
+        chatbot_ret += [(f"The time starts / est. ends:", f"{timestamp2string(start_time)} / {timestamp2string(end_time)}")]
+        chatbot_ret += [(f"The number of processed / est. total tokens:", f"{num_tokens_processed} / {num_tokens_total}")]
+        chatbot_ret += [(f"The dollars of processed / est. total tokens:", f"{dollars_tokens_processed} / {dollars_tokens_total}")]
+        
+    return chatbot_ret
+
+
+with gr.Blocks() as demo:
+    
+    with gr.Column(variant="panel"):
+        chatbot = gr.Chatbot()
+        txt = gr.Textbox(show_label=False, placeholder="Enter my API_KEY to access this private space").style(container=False)
+        with gr.Row():
+            button_showcase = gr.Button("Show Recent Query-Responses")
+            button_download = gr.Button("Download All Query-Responses")
+            button_display = gr.Button("Display Progress Infomation")
+    
+    downloadfile = gr.File(None, interactive=False, show_label=False, visible=False)
+    
+    button_showcase.click(fn=showcase, inputs=[txt], outputs=[chatbot])
+    button_download.click(fn=download, inputs=[txt], outputs=[chatbot, downloadfile])
+    button_display.click(fn=display, inputs=[txt], outputs=[chatbot])
+
+demo.launch()

dialogues_set/dialogues_film.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1065ddc9706d8644b5cb686e43c197e4ae1a6273a2ee85892ab014a3d0589048
+size 1990042

dialogues_set/dialogues_jindong.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c3cb0b5351f451d7079ba1f647c61d50c65955af1a231b4ebba5ad4336dfcec
+size 875228

dialogues_set/dialogues_music.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5edd415cdb29221cd727ea3536770440a56e0732f23e9dadc2037bc9c3242f2
+size 1316759

dialogues_set/dialogues_natural.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d94a0738b9d7778afa4294392a60bd4b4fd1da395c70bf7943c774da1cc9126a
+size 27800692

dialogues_set/dialogues_taobao.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4ce15361975a74e91acdd032862c7e3ac421d347057b1fb375647d5ac6bb826
+size 2472553

dialogues_set/dialogues_travel_kd.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c779ff0148f28403e3856b374fb7b4fac2125c749da067a55a7b499b7be2e87f
+size 1397624

requirements.txt

@@ -0,0 +1,2 @@
+openai==0.27.0
+tiktoken==0.3.0