Upload 8 files

.dockerignore

@@ -0,0 +1,9 @@
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+.venv/
+.env
+.git
+*.log
+news_data.json

Dockerfile

@@ -0,0 +1,18 @@
+# Dockerfile for deploying the FastAPI app on Hugging Face Spaces or any Docker environment
+FROM python:3.10-slim
+
+# Set working directory
+WORKDIR /app
+
+# Install Python dependencies
+COPY requirements.txt ./
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy application code
+COPY . .
+
+# Expose port (Hugging Face Spaces uses port 80)
+EXPOSE 80
+
+# Run the FastAPI app with Gunicorn and Uvicorn worker
+CMD ["gunicorn", "app:app", "--workers", "1", "--bind", "0.0.0.0:80", "--worker-class", "uvicorn.workers.UvicornWorker"]

agents.py

@@ -0,0 +1,102 @@
+import os
+import json
+import time
+import traceback
+
+from langchain_core.prompts import PromptTemplate
+from langchain_core.output_parsers import JsonOutputParser
+from pydantic import BaseModel, Field
+
+from langchain_groq import ChatGroq
+from dotenv import load_dotenv
+
+load_dotenv()
+
+analysis_llm = ChatGroq(
+    model="llama-3.1-8b-instant	",
+    temperature=0.8,
+    max_tokens=800,
+    timeout=None,
+    max_retries=2,
+    api_key=os.getenv("GROQ_ANALYSIS_API_KEY"),
+)
+post_content_llm = ChatGroq(
+    model="qwen-qwq-32b",
+    temperature=0.8,
+    timeout=None,
+    max_retries=2,
+    api_key=os.getenv("THREADS_POST_GENERATION_API_KEY"),
+)
+
+    
+def basic_analysis(news):
+    global analysis_llm_indx
+    prompt = PromptTemplate.from_file(
+        template_file="prompts/news_selector.yml",
+        input_variables=["news_object"],
+    )
+
+    for _ in range(5):
+        try:
+            response = analysis_llm.invoke(
+                prompt.invoke({"news_object": news})
+            )
+            print("################ BASIC ANALYSIS AGENT RESPONSE ################")
+            print(response.content)
+            print("################ BASIC ANALYSIS END AGENT RESPONSE ################")
+
+            start_index = response.content.find("{")
+            end_index = response.content.rfind("}")
+
+            print("start index:", start_index)
+            print("end index:", end_index)
+
+            abstracted_string = ""
+            if start_index != -1 and end_index != -1 and start_index < end_index:
+                abstracted_string = response.content[start_index : end_index + 1]
+
+                try:
+                    results = json.loads(abstracted_string)
+                    print(results)
+                    return results
+                except Exception as e:
+                    print(e)
+                    traceback.print_exc()
+        except Exception as e:
+            print(e)
+            traceback.print_exc()
+            time.sleep(30)
+
+    raise ValueError("LLM response is not in correct format.")
+
+
+def get_text_post_content(details, reference):
+    try:
+        prompt = PromptTemplate.from_file(
+            template_file="prompts/post_generator_without_source.yml",
+            input_variables=["NEWS_CONTENT", "CHAR_LENGTH"],
+        )
+
+        user_query = prompt.invoke(
+            {"NEWS_CONTENT": details, "CHAR_LENGTH": 490- len(reference)}
+        )
+        response = post_content_llm.invoke(user_query)
+
+        print("POST CONTENT RESPONSE:", response)
+        
+        content = response.content.replace('"', '')
+
+        if "</think>" in content:
+            content = content.split("</think>")[1]
+
+        start_indx = content.find("#")
+        content = f"""{content[:start_indx]}
+{reference}
+
+{content[start_indx:]}"""
+        
+        return content, True
+    except Exception as e:
+        print(e)
+        traceback.print_exc()
+        return "", False

app.py

@@ -0,0 +1,13 @@
+from fastapi import FastAPI
+
+from agents import basic_analysis, get_text_post_content
+
+app = FastAPI()
+
+@app.post("/get_basic_analysis")
+def handle_basic_analysis(news: dict):
+    return basic_analysis(news)
+
+@app.post("/get_text_post_content")
+def handle_text_post_content(details: str, reference: str):
+    return get_text_post_content(details, reference)

prompts/news_selector.yml

@@ -0,0 +1,27 @@
+input_variables:
+    - news_object
+    
+template: >-
+    You are an expert news analyzer. Your task is to analyze news to determine how it will impact the Indian stock market with the help of a title and a small description of the news if given. You will be given a news object in which some fields will be empty. You have to fill in those fields and return the same filled object.
+
+    Formate of news object:
+    {{
+        "title": "",
+        "url": "",
+        "description": " ",
+        "time": "",
+        "source": "",
+        "will_it_directly_impact_any_stock": "" -> boolean: true/false,
+        "directly_mentioned_companies_in_news": [] -> name of companies which are directly mentioned in the news and listed in indian stockmarket.,
+        "how_will_it_impact": "" -> categorical: Very Positive, Positive, Nautural, Negative, Very Negative.
+        "reason": "" -> reason for the impact.
+    }}
+
+    Output instructions:
+    - You have to strictly return news object after filling empty fields.
+    - You have to fill the fields in the same news object and return it.
+    - Don't give any additional information apart from filled news object.
+    - You must always return valid JSON fenced by a markdown code block. Do not return any additional text.
+
+    NEWS objects to analyze:
+    {news_object}

prompts/post_generator.yml

@@ -0,0 +1,26 @@
+# Not in use.
+input_variables:
+    - NEWS_CONTENT
+    - REFERENCE_URL
+    - CHAR_LENGTH
+    
+template: >-
+    You are an AI assistant specializing in creating concise, impactful social media posts for Threads. Your task is to generate a post based on the provided news and its reference link, ensuring it meets the following criteria:  
+
+    1. **Summary:** Provide a brief yet informative summary of the news, highlighting key details and numbers.  
+    2. **Impacted Stocks:** Identify Indian stocks that may be affected by this news but don't directly mention this in post content by saing impacted stocks.
+    3. **Character Limit:** Ensure the entire post does not exceed {CHAR_LENGTH} characters.
+    4. **Format:** You can use any format to present the news but above guidelines must be followed.
+    5. **HashTags:** Add few tranding hashtags under 20 chars in the post content.
+    6. **Source URL:** You must Include the full provided Source URL with given route in the post.  
+
+    Output instructions:
+    - Don't use emoji in the post content.
+    - You have to strictly return post content only not any other information or your thoughts.
+
+    Here is the news:  
+    {NEWS_CONTENT}  
+
+    Source URL: {REFERENCE_URL}  
+
+    Generate the post following these guidelines without add any additional details or information.  

prompts/post_generator_without_source.yml

@@ -0,0 +1,20 @@
+input_variables:
+    - NEWS_CONTENT
+    - CHAR_LENGTH
+    
+template: >-
+    You are an AI assistant specializing in creating concise, impactful social media posts for Threads. Your task is to generate a post based on the provided news and its reference link, ensuring it meets the following criteria:  
+
+    1. **Summary:** Provide a brief yet informative summary of the news, highlighting key details and numbers.  
+    2. **Impacted Stocks:** Identify Indian stocks that may be affected by this news but don't directly mention this in post content by saing impacted stocks.
+    3. **Character Limit:** Ensure the entire post does not exceed {CHAR_LENGTH} characters.
+    4. **Format:** You can use any format to present the news but above guidelines must be followed.
+    5. **HashTags:** Add only one related tranding hashtags under 10 chars at the last.
+
+    Output instructions:
+    - You have to strictly return post content only not any other information or your thoughts. Don't include Link in the post content. 
+
+    Here is the news:  
+    {NEWS_CONTENT}  
+
+    Generate the post following these guidelines without add any additional details or information.  

requirements.txt

@@ -0,0 +1,16 @@
+bs4==0.0.2
+html2text==2024.2.26
+pymongo[srv]==4.11
+selenium==4.28.1
+webdriver-manager==4.0.2
+tiktoken==0.8.0
+python-dotenv==1.0.1
+langchain==0.3.17
+langchain_core==0.3.33
+langchain_groq==0.2.4
+langchain_community==0.3.16
+fastapi===0.115.8
+gunicorn===23.0.0
+uvicorn===0.34.0
+pydantic==2.9.2
+certifi