Fix: moved large

.gitattributes

@@ -33,3 +33,5 @@ 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
+app/data/faiss_db/index.faiss filter=lfs diff=lfs merge=lfs -text
+app/data/faiss_db/index.pkl filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,68 @@
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from contextlib import asynccontextmanager
+
+from app.api import analysis, reports, chatbot, infos
+from langchain_community.chat_models import ChatOpenAI
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import FAISS
+from app.utils.rag import get_chain_disease, get_chain_infos, get_chain_chat
+import os, asyncio
+
+
+HUGGINGFACE_API_KEY = os.getenv("HUGGINGFACE_API_KEY")
+MISTRAL_API_KEY = os.getenv("MISTRAL_API_KEY")
+
+
+@asynccontextmanager
+async def lifespan(app: FastAPI):
+    embedding_model = await asyncio.to_thread(
+        HuggingFaceEmbeddings,
+        model_name="sentence-transformers/all-MiniLM-L6-v2",
+        model_kwargs={"use_auth_token": HUGGINGFACE_API_KEY}
+    )
+
+    vectorstore = await asyncio.to_thread(
+        FAISS.load_local,
+        r"app/data/faiss_db",
+        embeddings=embedding_model,
+        allow_dangerous_deserialization=True
+    )
+
+    llm = ChatOpenAI(
+        base_url="https://api.mistral.ai/v1",
+        api_key=MISTRAL_API_KEY,
+        model_name="mistral-medium"
+    )
+
+    app.state.diagnosis_chain = get_chain_disease(llm, vectorstore)
+    app.state.info_chain = get_chain_infos(llm, vectorstore)
+    app.state.chat_chain = get_chain_chat(llm=llm, vectorstore=vectorstore)
+
+    yield
+
+app = FastAPI(
+    title="Symptom Checker API",
+    description="API for symptom checking and analysis",
+    version="0.1.0",
+    lifespan=lifespan
+)
+
+# CORS configuration
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Include routers
+app.include_router(analysis.router, prefix="/api/analysis", tags=["analysis"])
+app.include_router(infos.router, prefix="/api/infos", tags=["infos"])
+app.include_router(reports.router, prefix="/api/reports", tags=["report"])
+app.include_router(chatbot.router, prefix="/api/chatbot", tags=["chat"])
+
+@app.get("/")
+async def root():
+    return {"message": "Symptom Checker API is running"}

app/api/analysis.py

@@ -0,0 +1,54 @@
+from fastapi import APIRouter, HTTPException,Request
+import json
+
+from app.utils.models import AnalysisResult, SymptomAnalysisRequest
+from typing import Union
+
+import re
+
+router = APIRouter()
+
+
+
+def process_rag_response(rag_response: Union[str, None]) -> AnalysisResult:
+    if not rag_response or not rag_response.strip():
+        raise ValueError("Empty response from RAG model")
+
+    # Extract JSON array from any extra text
+    match = re.search(r'\[\s*\{.*?\}\s*\]', rag_response, re.DOTALL)
+    if not match:
+        raise ValueError("No valid JSON array found in response")
+
+    try:
+        data = json.loads(match.group(0))
+    except json.JSONDecodeError as e:
+        raise ValueError(f"Invalid JSON format: {e}")
+
+    if not isinstance(data, list) or not data:
+        raise ValueError("Parsed JSON is not a valid non-empty list")
+
+    # Build the diagnoses dict
+    diagnoses = {
+        item.get("disease", "Unknown"): item.get("probability", 0) / 100
+        for item in data
+    }
+    
+    return AnalysisResult(diagnoses=diagnoses)
+
+
+
+def analyze_symptoms(request: Request, symptoms: str) -> AnalysisResult:
+    try:
+        rag_response = request.app.state.diagnosis_chain.run(symptoms)
+        return process_rag_response(rag_response)
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Symptom analysis failed: {str(e)}")
+
+
+@router.post("/analyze", response_model=AnalysisResult)
+async def analyze_selected_symptoms(request: Request, body: SymptomAnalysisRequest):
+    if not body.symptoms:
+        raise HTTPException(status_code=400, detail="At least one symptom is required")
+    return analyze_symptoms(request, body.symptoms)
+
+

app/api/chatbot.py

@@ -0,0 +1,24 @@
+from fastapi import  Request, APIRouter
+from app.utils.models import ChatRequest
+
+
+# Initialize FastAPI app
+router = APIRouter()
+
+# Store user queries
+chat_history_list = []
+
+
+
+# Endpoint for chatting
+@router.post("/chat")
+async def chat(request_chat: Request, request: ChatRequest):
+    user_question = request.question
+    chat_history_list.append(user_question)  # Track all queries
+    response = request_chat.app.state.chat_chain.run(user_question)
+
+    return {
+        "question": user_question,
+        "answer": response,
+        "history": chat_history_list
+    }

app/api/infos.py

@@ -0,0 +1,27 @@
+
+from app.utils.pdf_process import generate_pdf_report
+from app.utils.infos_process import extract_dict_from_text
+from app.utils.models import DiseaseInfo, DiseaseInput
+
+from fastapi import APIRouter, HTTPException, Request
+router = APIRouter()
+
+def get_disease_info(request: Request,disease_name: str) -> DiseaseInfo:
+    dict_str = request.app.state.info_chain.run(disease_name)
+    disease_info = extract_dict_from_text(dict_str)
+    return DiseaseInfo(infos=disease_info)
+
+
+
+
+@router.post("/infos", response_model=DiseaseInfo)
+async def give_full_infos(request: Request,requests: DiseaseInput):
+    if not requests.disease_name:
+        raise HTTPException(
+            status_code=400,
+            detail="At least one symptom is required"
+        )
+    disease_info=get_disease_info(request,requests.disease_name)
+    filename = f"app/report/report.pdf"
+    generate_pdf_report(disease_info.infos, filename)
+    return disease_info

app/api/reports.py

@@ -0,0 +1,21 @@
+from fastapi import APIRouter, HTTPException
+from fastapi.responses import FileResponse
+import os
+
+router = APIRouter()
+
+@router.get("/report")
+async def get_report():
+    report_path = f"app/report/report.pdf"
+    
+    if not os.path.exists(report_path):
+        raise HTTPException(
+            status_code=404,
+            detail="Report not found"
+        )
+    
+    return FileResponse(
+        report_path,
+        media_type="application/pdf",
+        filename=f"Report.pdf"
+    )

app/data/faiss_db/index.faiss

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f496ab9a89996bd7ed435529a7793d1844eccd8a11c4ac907bdebdeb2ade9868
+size 23780397

app/data/faiss_db/index.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b227f23738538e47fbda442e92c93b8736000cf42ec8e4f61af53ac2761e6b3b
+size 13678957

app/utils/infos_process.py

@@ -0,0 +1,48 @@
+import json
+import re
+
+def fix_multiline_strings(s):
+        lines = s.split('\n')
+        new_lines = []
+        buffer = ""
+        in_string = False
+
+        for line in lines:
+            line = line.rstrip()
+            quote_count = line.count('"')
+            if not in_string:
+                buffer = line
+                if quote_count % 2 == 1:  # Odd number of quotes -> string not closed
+                    in_string = True
+                else:
+                    new_lines.append(buffer)
+            else:
+                buffer += ' ' + line
+                if quote_count % 2 == 1:
+                    new_lines.append(buffer)
+                    in_string = False
+
+        if in_string:
+            new_lines.append(buffer)  # Last incomplete line
+        return "\n".join(new_lines)
+
+def extract_dict_from_text(text):
+    # Extract the code block
+    match = re.search(r"```(?:python)?\s*({.*?})\s*```", text, re.DOTALL)
+    if not match:
+        match = re.search(r"({.*})", text, re.DOTALL)
+    if not match:
+        return None
+
+    dict_str = match.group(1)
+
+    # Step 1: Replace unescaped line breaks inside strings
+    
+    dict_str = fix_multiline_strings(dict_str)
+
+    # Step 2: Try parsing as JSON
+    try:
+        return json.loads(dict_str)
+    except json.JSONDecodeError as je:
+        print("JSON decode error:", je)
+        return None

app/utils/map_details.py

@@ -0,0 +1,172 @@
+import requests
+from bs4 import BeautifulSoup
+from selenium import webdriver
+from selenium.webdriver.chrome.options import Options
+from selenium.webdriver.common.by import By
+from selenium.webdriver.support.ui import WebDriverWait
+from selenium.webdriver.support import expected_conditions as EC
+import geocoder
+import folium
+from folium import IFrame
+from geopy.distance import geodesic
+
+
+def get_doctor_links(city, speciality):
+    # Setup Chrome options
+    options = Options()
+    options.headless = True  # Run in headless mode
+    options.add_argument('--headless')  # Disable GPU acceleration
+    options.add_argument('--no-sandbox')
+    options.add_argument('--disable-dev-shm-usage')
+    # Initialize WebDriver (make sure chromedriver is in your PATH or specify path)
+    driver = webdriver.Chrome(options=options)
+    links = []
+    try:
+        url = f"https://www.med.tn/doctor/{speciality}/{city}"
+        driver.get(url)
+
+        # Wait for the doctor cards to load
+        WebDriverWait(driver, 10).until(
+            EC.presence_of_all_elements_located((By.CLASS_NAME, "card-doctor-block    "))
+        )
+
+        # Extract doctor cards
+        doctor_cards = driver.find_elements(By.CLASS_NAME, "card-doctor-block    ")
+        
+        for i, card in enumerate(doctor_cards, 1):
+            links.append(card.find_element(By.TAG_NAME, 'a').get_attribute('href'))
+
+    finally:
+        driver.quit()
+    return links
+
+
+
+def extract_doctor_profile(url):
+    # Headers to mimic a browser
+    headers = {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'
+    }
+
+    # Send request
+    response = requests.get(url, headers=headers)
+    response.raise_for_status()
+
+    # Parse HTML
+    soup = BeautifulSoup(response.text, 'html.parser')
+
+    # Extract profile label section
+    profile = soup.find('div', class_='profile__label')
+    if not profile:
+        raise Exception("Profile section not found")
+
+    # Extract image
+    img_tag = soup.find('div', class_='profile__photo')
+    img = ""
+    if img_tag and img_tag.find('img'):
+        img_src = img_tag.find('img')['src']
+        img = img_src.replace(
+            'https://imagecdn.med.ovh/unsafe/195x195/filters:format():quality(10):blur(30)/', '')
+
+    # Extract name, speciality, city
+    name = profile.find('h1').text.strip() if profile.find('h1') else ''
+    speciality = profile.find('div', class_='profile__label--spe')
+    speciality = speciality.text.strip() if speciality else ''
+    city = profile.find('div', class_='profile__label--adr')
+    city = city.text.strip() if city else ''
+
+    # Extract address
+    address_tag = soup.find('span', class_='profile__adr')
+    address = address_tag.text.strip() if address_tag else ''
+
+    # Extract phone numbers
+    phone_numbers = []
+    if soup.find('div', class_='displaynum'):
+        num_tels = soup.find('div', class_='displaynum') 
+        phone_numbers = [a.get_text(strip=True) for a in num_tels.find_all('a')]
+        if phone_numbers[-1].startswith('Book'):
+            phone_numbers.pop()
+    else:
+        phone_numbers = ['N/A']
+    # Extract map position
+    map_tag = soup.find('a', class_='btn-itineraire') or soup.find('a', target='_dir')
+    map_position = ""
+    if map_tag and 'href' in map_tag.attrs:
+        map_position = map_tag['href'].replace('?api=1&destination=', '')
+
+    # Final output
+    full_details = {
+        'img': img,
+        'name': name,
+        'speciality': speciality,
+        'city': city,
+        'address': address,
+        'phone_numbers': phone_numbers,
+        'map_position': map_position
+    }
+    return full_details
+
+def get_my_location():
+    g = geocoder.ip('me')
+    if not g.ok or not g.latlng:
+        raise RuntimeError("Unable to detect your location.")
+
+    user_lat, user_lon = g.latlng
+    city = g.city or "Your Location"
+    country = g.country or ""
+    return user_lat, user_lon, city, country
+
+def create_the_map(speciality):
+
+    # Get user location
+    user_lat, user_lon, city, country = get_my_location()
+    map = folium.Map(location=[user_lat, user_lon], zoom_start=12)
+    folium.Marker(
+        [user_lat, user_lon],
+        popup=f"You are here: {city}, {country}",
+        icon=folium.Icon(color='red', icon='user')
+    ).add_to(map)
+
+    # Fetch doctors
+    all_docs = get_doctor_links(city.lower(), speciality.lower())
+
+    for url in all_docs:
+        try:
+            doc = extract_doctor_profile(url)
+            if not doc.get("map_position"):
+                continue
+
+            lat_str, lon_str = doc['map_position'].replace('https://www.google.com/maps/dir/','').split(',')
+            lat, lon = float(lat_str), float(lon_str)
+
+            # Calculate distance to user
+            distance_km = geodesic((user_lat, user_lon), (lat, lon)).km
+            if distance_km > 4:
+                continue  # Skip doctors farther than 4km
+
+            # Build popup
+            img_html = f"<img src='{doc['img']}' width='100' height='100'><br>" if doc.get("img") else ""
+            phone_html = "<br>".join(doc.get("phone_numbers", []))
+            popup_html = f"""
+            {img_html}
+            <b>{doc['name']}</b><br>
+            <i>{doc['speciality']}</i><br>
+            <b>Address:</b> {doc['address']}<br>
+            <b>Phone:</b><br>{phone_html}
+            <br><b>Distance:</b> {distance_km:.2f} km
+            """
+
+            iframe = IFrame(popup_html, width=250, height=250)
+            popup = folium.Popup(iframe, max_width=300)
+
+            folium.Marker(
+                [lat, lon],
+                popup=popup,
+                icon=folium.Icon(color='blue', icon='plus')
+            ).add_to(map)
+
+        except Exception as e:
+            print(f"Error processing {url}: {e}")
+
+    # Save map
+    map.save("map.html")

app/utils/models.py

@@ -0,0 +1,21 @@
+from pydantic import BaseModel
+from typing import Optional
+
+class SymptomAnalysisRequest(BaseModel):
+    symptoms: str
+    language: Optional[str] = "en"
+    detailed: Optional[bool] = False
+
+# Pydantic model: only the diagnoses dictionary
+class AnalysisResult(BaseModel):
+    diagnoses: dict  # disease -> confidence (0.0 to 1.0)
+
+class DiseaseInfo(BaseModel):
+    infos: dict
+
+class DiseaseInput(BaseModel):
+    disease_name: str
+    language: Optional[str] = "en"
+
+class ChatRequest(BaseModel):
+    question: str

app/utils/pdf_process.py

@@ -0,0 +1,243 @@
+from reportlab.platypus import Frame, NextPageTemplate, PageTemplate, BaseDocTemplate, Paragraph, Image, Spacer,Paragraph
+from reportlab.lib.units import cm
+from reportlab.lib.styles import ParagraphStyle, getSampleStyleSheet
+from reportlab.lib.enums import TA_JUSTIFY
+from reportlab.lib.pagesizes import letter
+from reportlab.lib import colors
+from datetime import datetime
+
+def generate_pdf_report(disease_info, filename="report.pdf")->None:
+    """Generates a scientific article-style PDF report with two-column layout and a logo."""
+
+    # Custom styles
+    styles = getSampleStyleSheet()
+
+    # Add custom styles
+    styles.add(ParagraphStyle(
+        name='TitleStyle',
+        parent=styles['Heading1'],
+        fontSize=16,
+        leading=20,
+        alignment=1,  # Center
+        spaceAfter=20,
+        textColor=colors.darkblue
+    ))
+
+    styles.add(ParagraphStyle(
+        name='AuthorStyle',
+        parent=styles['Heading3'],
+        fontSize=10,
+        leading=12,
+        alignment=1,
+        spaceAfter=20,
+        textColor=colors.darkgrey
+    ))
+
+    styles.add(ParagraphStyle(
+        name='AbstractStyle',
+        parent=styles['BodyText'],
+        fontSize=10,
+        leading=12,
+        alignment=TA_JUSTIFY,
+        backColor=colors.lightgrey,
+        borderPadding=5,
+        spaceAfter=20
+    ))
+
+    styles.add(ParagraphStyle(
+        name='SectionHeader',
+        parent=styles['Heading2'],
+        fontSize=12,
+        leading=14,
+        spaceBefore=15,
+        spaceAfter=10,
+        textColor=colors.darkblue,
+        underlineWidth=1,
+        underlineColor=colors.darkblue,
+        underlineOffset=-5
+    ))
+
+    styles.add(ParagraphStyle(
+        name='LeftColumn',
+        parent=styles['BodyText'],
+        fontSize=10,
+        leading=12,
+        alignment=TA_JUSTIFY,
+        leftIndent=0,
+        rightIndent=5,
+        spaceAfter=7
+    ))
+
+    styles.add(ParagraphStyle(
+        name='RightColumn',
+        parent=styles['BodyText'],
+        fontSize=10,
+        leading=12,
+        alignment=TA_JUSTIFY,
+        leftIndent=5,
+        rightIndent=0,
+        spaceAfter=7
+    ))
+
+    styles.add(ParagraphStyle(
+        name='BulletPoint',
+        parent=styles['BodyText'],
+        fontSize=10,
+        leading=12,
+        leftIndent=15,
+        bulletIndent=0,
+        spaceAfter=3,
+        bulletFontName='Symbol',
+        bulletFontSize=8
+    ))
+
+    styles.add(ParagraphStyle(
+        name='Reference',
+        parent=styles['Italic'],
+        fontSize=8,
+        leading=10,
+        textColor=colors.darkgrey,
+        spaceBefore=15
+    ))
+
+    # Create document with two columns
+    class TwoColumnDocTemplate(BaseDocTemplate):
+        def __init__(self, filename, **kw):
+            BaseDocTemplate.__init__(self, filename, **kw)
+            # Calculate column widths
+            page_width = self.pagesize[0] - 2*self.leftMargin
+            col_width = (page_width - 1*cm) / 2  # 1cm gutter
+
+            # First page template with title
+            first_page = PageTemplate(id='FirstPage',
+                frames=[
+                    Frame(self.leftMargin, self.bottomMargin, 
+                          col_width, self.height, 
+                          id='leftCol'),
+                    Frame(self.leftMargin + col_width + 1*cm, 
+                          self.bottomMargin, 
+                          col_width, self.height, 
+                          id='rightCol')
+                ])
+            self.addPageTemplates(first_page)
+
+            # Other pages template
+            other_pages = PageTemplate(id='OtherPages',
+                frames=[
+                    Frame(self.leftMargin, self.bottomMargin, 
+                          col_width, self.height, 
+                          id='leftCol2'),
+                    Frame(self.leftMargin + col_width + 1*cm, 
+                          self.bottomMargin, 
+                          col_width, self.height, 
+                          id='rightCol2')
+                ])
+            self.addPageTemplates(other_pages)
+
+    doc = TwoColumnDocTemplate(filename,
+                             pagesize=letter,
+                             leftMargin=2*cm,
+                             rightMargin=2*cm,
+                             topMargin=2*cm,
+                             bottomMargin=2*cm)
+
+    story = []
+
+    # Add logo at the top (centered)
+    try:
+        logo = Image('app/data/logo_platform.jpg', width=6*cm, height=2*cm)
+        logo.hAlign = 'CENTER'
+        story.append(logo)
+        story.append(Spacer(1, 0.3*cm))
+    except Exception as e:
+        pass  # If logo not found, skip
+
+    # Title and authors
+    title = Paragraph(disease_info.get('Title', 'Medical Condition Report'), styles['TitleStyle'])
+    authors = Paragraph("Generated by AIHealthCheck AI Assistant", styles['AuthorStyle'])
+    date = Paragraph(datetime.now().strftime("%B %d, %Y"), styles['AuthorStyle'])
+
+    story.append(title)
+    story.append(authors)
+    story.append(date)
+    story.append(NextPageTemplate('OtherPages'))  # Switch to two-column layout
+
+    # Abstract
+    abstract_text = f"<b>Abstract</b><br/><br/>{disease_info.get('Overview', 'No overview available.')}"
+    abstract = Paragraph(abstract_text, styles['AbstractStyle'])
+    story.append(abstract)
+
+    # Function to format content
+
+    def format_content(text, styles):
+        if not text:
+            return []
+
+        # If it's a list, return each item as a bullet paragraph
+        if isinstance(text, list):
+            return [Paragraph(f"• {item}", styles) for item in text]
+
+        # If it's a dict, render each key-value pair
+        if isinstance(text, dict):
+            items = []
+            for key, value in text.items():
+                if isinstance(value, list):
+                    items.append(Paragraph(f"<b>{key}:</b>", styles))
+                    items.extend([Paragraph(f"• {v}", styles) for v in value])
+                else:
+                    items.append(Paragraph(f"• <b>{key}:</b> {value}", styles))
+            return items
+
+        # Otherwise, treat it as a simple paragraph
+        return [Paragraph(text, styles)]
+
+
+
+    # Organize content into left and right columns
+    left_column_content = [
+        ('Symptoms', disease_info.get('Symptoms')),
+        ('Causes', disease_info.get('Causes')),
+        ('Risk Factors', disease_info.get('Risk factors')),
+        ('Complications', disease_info.get('Complications'))
+    ]
+
+    right_column_content = [
+        ('Diagnosis', disease_info.get('Diagnosis')),
+        ('Treatment', disease_info.get('Treatment')),
+        ('Prevention', disease_info.get('Prevention')),
+        ('When to See a Doctor', disease_info.get('When to see a doctor')),
+        ('Lifestyle and Home Remedies', disease_info.get('Lifestyle and home remedies'))
+    ]
+
+    # Add left column content
+    for section, content in left_column_content:
+        if content:
+            story.append(Paragraph(section, styles['SectionHeader']))
+            formatted = format_content(content, styles['LeftColumn'])
+            if isinstance(formatted, list):
+                story.extend(formatted)
+            else:
+                story.append(formatted)
+
+    # Switch to right column
+    story.append(NextPageTemplate('OtherPages'))
+
+    # Add right column content
+    for section, content in right_column_content:
+        if content:
+            story.append(Paragraph(section, styles['SectionHeader']))
+            formatted = format_content(content, styles['RightColumn'])
+            if isinstance(formatted, list):
+                story.extend(formatted)
+            else:
+                story.append(formatted)
+
+    # Add references
+    story.append(Paragraph("Medical Recommendations", styles['SectionHeader']))
+    story.append(Paragraph(disease_info.get('Medical Recommendation', 'No medical recommendations available.'), styles['LeftColumn']))
+    story.append(Paragraph("References", styles['SectionHeader']))
+    story.append(Paragraph("1. Mayo Clinic Medical References", styles['Reference']))
+
+    # Build the PDF
+    doc.build(story)
+

app/utils/rag.py

@@ -0,0 +1,128 @@
+
+from langchain.prompts import PromptTemplate
+from langchain.chains import RetrievalQA
+from langchain.prompts import ChatPromptTemplate
+import warnings
+from langchain.memory import ConversationBufferMemory
+from langchain.chains import ConversationalRetrievalChain, LLMChain, StuffDocumentsChain
+warnings.filterwarnings("ignore")
+import os
+
+
+def get_chain_infos(llm, vectorstore):
+# Prompt templates
+    prompt_template = ChatPromptTemplate.from_template("""
+You are a knowledgeable and precise medical assistant. Your task is to extract and summarize key information 
+from the raw disease content into a structured, clean, and medically accurate Python dictionary format with the following keys:
+['Title', 'Overview', 'Symptoms', 'Causes', 'Risk factors', 'Complications', 'Prevention',
+ 'When to see a doctor', 'Diagnosis', 'Treatment', 'Lifestyle and home remedies', 'Medical Recommendation']
+
+Disease Raw Info:
+{context}
+
+Guidelines:
+- Provide concise, factual, and medically sound content for each field.
+- For 'Medical Recommendation':
+  • Give the most accurate and actionable advice based on the disease context.
+  • If medical consultation is required, specify the **exact medical specialty** (e.g., cardiologist, neurologist).
+  • If home care is sufficient, describe **precise steps or remedies** clearly and briefly.
+  • Keep it **tailored, specific, and informative**—avoid vague generalities.
+  • When applicable, be brief without sacrificing clarity or accuracy.
+
+Output:
+Return only a valid JSON-like Python dictionary containing the summarized disease information under each specified key. Do not include any extra commentary or explanation.
+""")
+    # Initialize the RetrievalQA chain
+    info_chain = RetrievalQA.from_chain_type(
+    llm=llm,
+    retriever=vectorstore.as_retriever(search_kwargs={"k": 1}),
+    chain_type="stuff",  # or "map_reduce", etc. depending on what you want
+    chain_type_kwargs={"prompt": prompt_template},
+    return_source_documents=False
+)
+
+    return info_chain
+
+def get_chain_disease(llm, vectorstore):
+
+    prompt = PromptTemplate(
+    input_variables=["context", "question"],
+    template="""
+You are a medical diagnosis assistant.
+
+Use the context below to identify which disease best matches the given symptoms.
+
+Return your response in this format as a JSON array of objects WITHOUT any additional text:
+[
+  {{
+    "disease": "Disease Name",
+    "probability": 87
+  }},
+  {{
+    "disease": "Other Likely Disease",
+    "probability": 13
+  }}
+]
+
+### Context:
+{context}
+
+### Symptoms:
+{question}
+"""
+)
+
+# RAG Chains
+    diagnosis_chain = RetrievalQA.from_chain_type(
+    llm=llm,
+    retriever=vectorstore.as_retriever(search_kwargs={"k": 3}),
+    chain_type_kwargs={"prompt": prompt},
+    return_source_documents=False
+)
+    return diagnosis_chain
+
+def get_chain_chat(llm, vectorstore):
+    # Memory
+    memory = ConversationBufferMemory(
+        memory_key="chat_history",
+        return_messages=True
+    )
+
+    # Prompts
+    answer_prompt_template = ChatPromptTemplate.from_template("""
+You are a bilingual knowledgeable and precise medical assistant fluent in both English and French. Your task is to provide accurate and concise answers to medical queries.
+Detect the language of the query (English or French) and respond in the same language.
+When answering, please ensure that your response is clear with no extra explanation. If the question is ambiguous or requires further clarification, ask for more details.
+
+Context: {context}
+Question: {question}
+Answer:
+""")
+
+    qa_llm_chain = LLMChain(llm=llm, prompt=answer_prompt_template)
+
+    combine_docs_chain = StuffDocumentsChain(
+        llm_chain=qa_llm_chain,
+        document_variable_name="context"
+    )
+
+    condense_question_prompt = PromptTemplate.from_template("""
+Given the following conversation and a follow-up question, rephrase the follow-up question to be a standalone question.
+
+Chat History:
+{chat_history}
+Follow-Up Input: {question}
+Standalone question:
+""")
+
+    question_generator = LLMChain(llm=llm, prompt=condense_question_prompt)
+
+    # Conversational chain
+    chat_chain = ConversationalRetrievalChain(
+        retriever=vectorstore.as_retriever(search_kwargs={"k": 2}),
+        memory=memory,
+        question_generator=question_generator,
+        combine_docs_chain=combine_docs_chain,
+        return_source_documents=False
+    )
+    return chat_chain

report/report.pdf

@@ -0,0 +1,86 @@
+%PDF-1.4
+%���� ReportLab Generated PDF document http://www.reportlab.com
+1 0 obj
+<<
+/F1 2 0 R /F2 3 0 R /F3 4 0 R /F4 5 0 R
+>>
+endobj
+2 0 obj
+<<
+/BaseFont /Helvetica /Encoding /WinAnsiEncoding /Name /F1 /Subtype /Type1 /Type /Font
+>>
+endobj
+3 0 obj
+<<
+/BaseFont /Helvetica-Bold /Encoding /WinAnsiEncoding /Name /F2 /Subtype /Type1 /Type /Font
+>>
+endobj
+4 0 obj
+<<
+/BaseFont /Helvetica-BoldOblique /Encoding /WinAnsiEncoding /Name /F3 /Subtype /Type1 /Type /Font
+>>
+endobj
+5 0 obj
+<<
+/BaseFont /Helvetica-Oblique /Encoding /WinAnsiEncoding /Name /F4 /Subtype /Type1 /Type /Font
+>>
+endobj
+6 0 obj
+<<
+/Contents 10 0 R /MediaBox [ 0 0 612 792 ] /Parent 9 0 R /Resources <<
+/Font 1 0 R /ProcSet [ /PDF /Text /ImageB /ImageC /ImageI ]
+>> /Rotate 0 /Trans <<
+
+>> 
+  /Type /Page
+>>
+endobj
+7 0 obj
+<<
+/PageMode /UseNone /Pages 9 0 R /Type /Catalog
+>>
+endobj
+8 0 obj
+<<
+/Author (\(anonymous\)) /CreationDate (D:20250524213346+01'00') /Creator (\(unspecified\)) /Keywords () /ModDate (D:20250524213346+01'00') /Producer (ReportLab PDF Library - www.reportlab.com) 
+  /Subject (\(unspecified\)) /Title (\(anonymous\)) /Trapped /False
+>>
+endobj
+9 0 obj
+<<
+/Count 1 /Kids [ 6 0 R ] /Type /Pages
+>>
+endobj
+10 0 obj
+<<
+/Filter [ /ASCII85Decode /FlateDecode ] /Length 2091
+>>
+stream
+Gatm<9lo&I&A@sBlqtkq%)Hj'*BU5Vj"tYJf$bLD?kO)LPCs"!9%75*"\D94D+RQlO_n5GEo4cTnTGu[f_?Z8hkU6s5DOG)#nR@($Qu&&_>G'4bV-\;ORC9,Z%CVb$&!J<\TUO555\T6C&+?=`t?DS"i_^tckrRoYRZa[LNRDR1if]XNa`g>T;OC=9QM[DI09(j@/^pRrND@"+8C/971b4"=BIlIKUr^lpE:cZA`NotRtWnl&]'hufeN>_Pt]V`V=;!Gb>1'Uk#HSiZ/c3U_51%7<OAiNN;L(+@I3g)MK*0:P/ssLDSYe>Qsdtl-aIs@@9UD$UD7JY"GZA9X+ZnEIiR(JM!N&G>T4r@`S2_YO#>=$MiMXQ!=aCfH-%@+/g"(>C!pUuWP%?pc!:9PETjn8k!+"m!aGb['pVbgMC4L+`Md6r;+ajX8jBL/PG2AqgD&<:IObafa*!hI<Q<sZ=%n0I!&#->d<?<lIBRoBR'PXKV/]P!X_H'Qr)!*=Gue<C@G"#4@`9"L""dGJp#GJHA4nOI22-_7MO-uWpjeTE7@kNEDA-[!ekEM<1A[fM:GSO)aQBcZCRXFU.?oe6*$93;&)n.I(T5'<&/S3/?=B#HEXG$UQ%@9^So%8?gP1s?.T9A$Napacj99uGk71RSCY,8)RH/LJd<MWE=ZkXlRoPWZ]*!OU<iKdd<++CN]+)156EM-$Z<h!4oJ.lNhgCh`QAD8CA]iG-JB=6,#qrQdKPWE+8uA"5=e&'dSq&#;r<a%$[MKQDV/g`OeHqnOgW?p!Z)U3c%$8,q-JF5/88fkO*_7-b%5&*Y3\Z))m4hGpH8T:77ts#d$go7App+M0EST1aPsIl<RUda1TfWJBGMLZ+)Sl.)(mr&@C)T8`"!H2;"8'-)8^:.7pJQ2<b77!YW%4E$K;*DgK,P8,>Ej%0`83!9Vaf[ROC9NObGcsZb3NNk78VSW-HoXs;n.1JLOQIeZClK[nWpc5f>3oZl-G,D]uICurcD;#<Q=V.nldp*G*fZ-(l/t$]hpfhWYP(L[T4O\1/6ctA>Y4X>,FI+$)9HgRT28BFN`9>=\u+q\M66_QKul,p2+SmpT.*_e("%a"2\J'VVH)=NcE_MGE6;]7F`01!Ob3bjfu'C93NY]"!hfLaT!i,9KZT!a<>Oml2*W=:H1#*^l+r,VQDbV@`<'6<pf;LIk-@(<]QICAQC[^4bPo#C$O9nE-Pdp5HOgd'3HNu=Yh>$2q8MI/&Zau3(Og5DfcgMLdu<>060WW=AYU@no2_;dDAc8i*$YaW4[KYFkI4f%rVkarSa>-Y@Odr?'YQJRfO3@p:9,hr<''Udr3kVQhS$%kXh/AFIt",,)&Y-C89<VdM5R*__d=ZSS9_2B;$&/Pc'a[FaCH@*UY5cFnj@g&C4#&>X'7!d45;In!&k-#?nPN\RFj6*p+*[6<Eh^r[cmnn6Mfd''4hBX5`7-&WKQC4.og6fe[^dO1]:H7q7@8IYBs8a`8O..Y#Q^$ce,@;B&-"0Bjj6RqGPJQ>i?/)Mb2DJ6*\Rglp\^HX?.72XE8d(S>FYrimL>-i3&H&K`oV"rps:Xg:"BiE21YEiKFRWa@C]!uCkb)f_*r2PtSM1Vm6:%!3j-Y8L4?-LN=TEL[=rIQ]i2k7b5L8M7%0Qm^NrP>gW(./&Fd(X:[email protected](pPkg6)!Lr]W]YO8JGd_gW!F6I2n@W:$XNFARBVi,'bB0q\(\K*@$gm0YKOMDK^**jPe>F+jt:b2h6.BXPm)S2bQ_7lo,E@o'Mj0CjFTo-Z[D[:]h3(RWNd9\'TK"5hH@hb5,R'REF1L,;^IQ],^C[sWatBToblmVYG/i%*953+CWci(eB/$oqL7?P0i"br2PsB$09>"qU\ejQ0S7VRgu#[email protected]>`Je?^l12[UQd2G;RG7Mn/d[S&c@)AsS:qF!^AJETkW1CCd-D[Q48X&.@o[bYr29e_AK8s4b!HLC*6!p>tW.D"jB!<&]()L:u_*MP4OrB=\[oL#s-Eh<.N9cH;LJbOXauBEq./$IuMUD.]pR:<oRsH@C$Dq_5Rl^3VM,_o7*`\QR#~>endstream
+endobj
+xref
+0 11
+0000000000 65535 f 
+0000000073 00000 n 
+0000000134 00000 n 
+0000000241 00000 n 
+0000000353 00000 n 
+0000000472 00000 n 
+0000000587 00000 n 
+0000000781 00000 n 
+0000000849 00000 n 
+0000001132 00000 n 
+0000001191 00000 n 
+trailer
+<<
+/ID 
+[<6b5b3613f29d0411c75de7a4437a5a87><6b5b3613f29d0411c75de7a4437a5a87>]
+% ReportLab generated PDF document -- digest (http://www.reportlab.com)
+
+/Info 8 0 R
+/Root 7 0 R
+/Size 11
+>>
+startxref
+3374
+%%EOF

requirements.txt

@@ -0,0 +1,13 @@
+fastapi 
+pydantic<2.0.0
+uvicorn
+langchain 
+faiss-cpu 
+sentence-transformers 
+openai
+reportlab
+geopy
+folium
+geocoder
+selenium
+bs4