Update main.py

main.py

@@ -8,7 +8,6 @@ from PyPDF2 import PdfReader
 from PIL import Image
 import fitz  # PyMuPDF
 import openai
-import pytesseract
 from dotenv import load_dotenv
 
 # Load environment variables
@@ -28,7 +27,6 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-
 def vision(file_content):
     """Extract text from images inside a PDF using PyMuPDF & OCR."""
     pdf_document = fitz.open(stream=file_content, filetype="pdf")
@@ -65,7 +63,6 @@ def vision(file_content):
     except Exception as e:
         raise HTTPException(status_code=500, detail=f"Error in GPT-4o vision processing: {str(e)}")
 
-
 @app.post("/get_ocr_data/")
 def get_data(input_file: UploadFile = File(...)):
     """Extract structured data from a PDF resume."""
@@ -77,7 +74,7 @@ def get_data(input_file: UploadFile = File(...)):
 
         if file_type == "application/pdf":
             pdf_reader = PdfReader(io.BytesIO(file_content))
-            
+
             for page in pdf_reader.pages:
                 text = page.extract_text()
                 if text:
@@ -86,36 +83,36 @@ def get_data(input_file: UploadFile = File(...)):
             if not extracted_text.strip():  # If no text found, use vision processing
                 print("\nVision OCR running...\n")
                 extracted_text = vision(file_content)
-
         else:
             raise HTTPException(status_code=400, detail="Unsupported file type")
 
         print("Extracted Text:\n", extracted_text.strip())
 
-        # Call GPT-4o to structure extracted text into JSON format
-        prompt = f"""This is CV data: {extracted_text.strip()}.
-        IMPORTANT: The output should be a JSON array! Make sure the JSON is valid.
-        If no data is found, fill missing fields with "none". Do not include extra explanation text.
+        # Call GPT-4o to process extracted text into structured JSON
+        prompt = f"""
+        This is CV data: {extracted_text.strip()}.
+        IMPORTANT: The output should be a JSON array! Make sure the JSON is valid. If no data is found, fill missing fields with "none". Do not add any extra explanation text.
+        Need only JSON output.
         
         Example Output:
         ```json
-        {{
-           "firstname": "First Name",
-           "lastname": "Last Name",
-           "email": "Email Address",
-           "contact_number": "Contact Number",
-           "home_address": "Full Home Address",
-           "home_town": "Home Town or City",
-           "total_years_of_experience": "Total Years of Experience",
-           "education": "Institution Name, Degree Name",
-           "LinkedIn_link": "LinkedIn URL",
-           "experience": "Job Title, Start Date - End Date, Company Name; Job Title, Start Date - End Date, Company Name; Job Title, Start Date - End Date, Company Name",
+        [
+           "firstname": "firstname",
+           "lastname": "lastname",
+           "email": "email",
+           "contact_number": "contact number",
+           "home_address": "full home address",
+           "home_town": "home town or city",
+           "total_years_of_experience": "total years of experience",
+           "education": "Institution Name, Country, Degree Name, Graduation Year; Institution Name, Country, Degree Name, Graduation Year",
+           "LinkedIn_link": "LinkedIn link",
+           "experience": "experience",
            "industry": "industry of work",
-           "skills": "Skill 1, Skill 2, Skill 3",
-           "positions": ["Job Title 1", "Job Title 2"],
-           "summary": "Summary of qualifications and experience"
-        }}
-        ```"""
+           "skills": "skills (Identify and list specific skills mentioned in both the skills section and inferred from the experience section), formatted as: Skill 1, Skill 2, Skill 3, Skill 4, Skill 5",
+           "positions": ["Job title 1, Job title 2, Job title 3"]
+        ]
+        ```
+        """
 
         response = openai.ChatCompletion.create(
             model="gpt-4o",
@@ -133,9 +130,3 @@ def get_data(input_file: UploadFile = File(...)):
 
     except Exception as e:
         raise HTTPException(status_code=500, detail=f"Error processing file: {str(e)}")
-
-      
-   
-   
-
-