Update app.py

app.py

@@ -1,64 +1,84 @@
-import requests
 import streamlit as st
+import requests
+import fitz  # PyMuPDF for PDF extraction
 from sentence_transformers import SentenceTransformer, util
+import torch
+
+# -------------------- CONFIG --------------------
+st.set_page_config(page_title="AI Job Matcher", page_icon="🤖", layout="wide")
 
-# Load embedding model
 @st.cache_resource
 def load_model():
     return SentenceTransformer("all-MiniLM-L6-v2")
 
 model = load_model()
 
-# Fetch jobs from RemoteOK
+# -------------------- FUNCTIONS --------------------
+def extract_text_from_pdf(uploaded_file):
+    """Extract text from uploaded PDF."""
+    text = ""
+    doc = fitz.open(stream=uploaded_file.read(), filetype="pdf")
+    for page in doc:
+        text += page.get_text("text")
+    return text
+
 def fetch_jobs():
+    """Fetch jobs from RemoteOK API."""
     url = "https://remoteok.com/api"
     try:
         response = requests.get(url, headers={"User-Agent": "Mozilla/5.0"})
-        data = response.json()
-        jobs = data[1:15]  # skip metadata, get first 15 jobs
-        return jobs
+        if response.status_code == 200:
+            return response.json()[1:]  # skip metadata
     except Exception as e:
-        st.error(f"Error fetching jobs: {e}")
-        return []
+        st.error(f"⚠️ Error fetching jobs: {e}")
+    return []
 
-# Match jobs with user skills
-def match_jobs(user_skills):
-    jobs = fetch_jobs()
-    if not jobs:
+def match_jobs(cv_text, jobs, top_n=5):
+    """Match CV text with job postings using embeddings."""
+    if not cv_text.strip():
         return []
 
-    user_embedding = model.encode(user_skills, convert_to_tensor=True)
-
+    cv_embedding = model.encode(cv_text, convert_to_tensor=True)
     results = []
-    for job in jobs:
-        desc = f"{job.get('position','')} at {job.get('company','')}. Tags: {job.get('tags','')}"
-        job_embedding = model.encode(desc, convert_to_tensor=True)
-        similarity = util.cos_sim(user_embedding, job_embedding).item()
+
+    for job in jobs[:30]:  # limit for demo
+        title = job.get("position", "")
+        company = job.get("company", "")
+        desc = job.get("description", "")
+        job_text = f"{title} at {company}. {desc}"
+
+        job_embedding = model.encode(job_text, convert_to_tensor=True)
+        similarity = util.cos_sim(cv_embedding, job_embedding).item()
+
         results.append({
-            "company": job.get("company"),
-            "position": job.get("position"),
-            "url": job.get("url"),
+            "title": title,
+            "company": company,
+            "url": job.get("url", ""),
             "score": round(similarity * 100, 2)
         })
 
-    results = sorted(results, key=lambda x: x["score"], reverse=True)
-    return results[:5]
-
-# ---------------- UI ----------------
-st.set_page_config(page_title="AI Freelancer Job Matcher", page_icon="🤖", layout="wide")
+    return sorted(results, key=lambda x: x["score"], reverse=True)[:top_n]
 
+# -------------------- UI --------------------
 st.title("🤖 AI Freelancer Job Matcher")
-st.markdown("Find jobs in real-time from **RemoteOK**, matched to your skills using AI.")
+st.markdown("Upload your **CV (PDF)** and get real-time job matches from RemoteOK.")
 
-skills = st.text_input("Enter your skills (comma separated)", "Python, AI, Docker")
+uploaded_file = st.file_uploader("📄 Upload your CV", type=["pdf"])
+
+if uploaded_file is not None:
+    cv_text = extract_text_from_pdf(uploaded_file)
+    st.success("✅ CV uploaded & text extracted!")
+    st.text_area("Extracted CV Text", cv_text[:1000], height=200)
+
+    jobs = fetch_jobs()
+    if jobs:
+        results = match_jobs(cv_text, jobs)
 
-if st.button("Find Jobs"):
-    results = match_jobs(skills)
-    if results:
-        for job in results:
-            st.subheader(f"🔹 {job['position']} at {job['company']}")
-            st.write(f"**Match:** {job['score']}%")
-            st.markdown(f"[Apply here]({job['url']})")
+        st.subheader("🎯 Top Job Matches")
+        for r in results:
+            st.markdown(f"**{r['title']}** at *{r['company']}*")
+            st.write(f"Match Score: {r['score']}%")
+            st.markdown(f"[View Job Posting]({r['url']})")
             st.markdown("---")
     else:
-        st.warning("No jobs found. Try again later.")
+        st.warning("⚠️ No jobs fetched. Try again later.")