Update app.py

app.py

@@ -1,72 +1,95 @@
 import streamlit as st
-from backend.agents.remoteok_agent import fetch_remoteok_jobs
-from backend.agents.matcher import compute_match, extract_keywords
-from backend.agents.resume_parser import extract_text
-from bs4 import BeautifulSoup
-import re
-
-st.set_page_config(page_title="MATCHHIVE", layout="wide")
-st.title("🚀 MATCHHIVE - AI Job Matcher")
-
-# Function to clean HTML
-def clean_html(raw_html):
-    soup = BeautifulSoup(raw_html, "html.parser")
-    # Replace <br> and <p> with newlines
-    for br in soup.find_all(["br", "p"]):
-        br.insert_after("\n")
-    # Extract text and remove extra whitespace
-    text = soup.get_text(separator="\n")
-    lines = [line.strip() for line in text.splitlines() if line.strip()]
-    return "\n".join(lines)
-
-# Upload resume
-uploaded = st.file_uploader("Upload your resume (PDF/DOCX)", type=["pdf", "docx"])
-resume_text = None
-
-if uploaded:
-    with open("temp_resume." + uploaded.name.split(".")[-1], "wb") as f:
-        f.write(uploaded.read())
-    resume_text = extract_text(f.name)
-    st.success("✅ Resume uploaded & parsed!")
-
-    # Extract keywords/skills from resume
-    resume_keywords = extract_keywords(resume_text)
-
-    # Fetch jobs
-    st.subheader("Job Feed - Matched Jobs")
-    jobs = fetch_remoteok_jobs(limit=10)
-    matched_jobs = []
-
-    for job in jobs:
-        # Clean job description from HTML
-        clean_desc = clean_html(job["description"])
-
-        # Compute match score
-        score = compute_match(resume_text, clean_desc)
-        if score > 0.1:  # show only jobs with some match
-            matched_jobs.append((job, score, clean_desc))
-
-    if matched_jobs:
-        for job, score, clean_desc in matched_jobs:
-            st.markdown(f"### {job['title']} at {job['company']}")
-
-            # Highlight matched keywords
-            highlighted_desc = clean_desc
-            for kw in resume_keywords:
-                # Use regex to highlight words ignoring case
-                highlighted_desc = re.sub(
-                    fr"(?i)\b({re.escape(kw)})\b",
-                    r"**\1**",
-                    highlighted_desc
-                )
-
-            st.markdown(highlighted_desc[:500] + "...")  # truncate for preview
-            st.progress(score)
-            st.write(f"**Match Score:** {score:.2f}")
-            st.markdown(f"[Apply Here]({job['url']})")
-            st.divider()
-    else:
-        st.info("No matching jobs found for your resume yet.")
-
-else:
-    st.info("Please upload your resume to see relevant job offers.")
+import requests
+from sentence_transformers import SentenceTransformer
+import faiss
+import numpy as np
+from groq import Groq
+
+# -----------------------------
+# CONFIG
+# -----------------------------
+REMOTEOK_URL = "https://remoteok.com/api"
+EMBED_MODEL = "BAAI/bge-small-en-v1.5"
+
+# Load embedding model
+@st.cache_resource
+def load_model():
+    return SentenceTransformer(EMBED_MODEL)
+
+model = load_model()
+
+# Initialize Groq client
+groq_client = Groq(api_key=st.secrets.get("GROQ_API_KEY", None))
+
+# -----------------------------
+# FUNCTIONS
+# -----------------------------
+
+def fetch_jobs():
+    resp = requests.get(REMOTEOK_URL)
+    if resp.status_code == 200:
+        jobs = resp.json()[1:]  # skip metadata
+        return jobs
+    return []
+
+def embed_texts(texts):
+    return model.encode(texts, convert_to_numpy=True, normalize_embeddings=True)
+
+def match_jobs(resume_text, jobs, top_k=5):
+    # prepare job descriptions
+    job_texts = [f"{job.get('position','')} {job.get('company','')} {job.get('description','')}" for job in jobs]
+    
+    # embeddings
+    resume_vec = embed_texts([resume_text])
+    job_vecs = embed_texts(job_texts)
+
+    # FAISS index
+    dim = job_vecs.shape[1]
+    index = faiss.IndexFlatIP(dim)  # cosine similarity (normalized)
+    index.add(job_vecs)
+
+    scores, idx = index.search(resume_vec, top_k)
+    results = []
+    for i, score in zip(idx[0], scores[0]):
+        results.append((jobs[i], float(score)))
+    return results
+
+def generate_resume(resume_text, job):
+    prompt = f"""
+    You are an AI career assistant. 
+    Given this resume:\n{resume_text}\n
+    and this job description:\n{job['description']}\n
+    Generate a tailored one-page resume that highlights relevant skills and experience.
+    Keep it concise and professional.
+    """
+
+    chat_completion = groq_client.chat.completions.create(
+        model="llama-3.1-70b-versatile",
+        messages=[{"role": "user", "content": prompt}],
+        temperature=0.7,
+    )
+    return chat_completion.choices[0].message["content"]
+
+# -----------------------------
+# STREAMLIT UI
+# -----------------------------
+st.title("MATCHHIVE - AI Job Matcher")
+
+resume_file = st.file_uploader("Upload your resume (txt/pdf/docx)", type=["txt"])
+if resume_file:
+    resume_text = resume_file.read().decode("utf-8", errors="ignore")
+
+    st.subheader("Fetching jobs...")
+    jobs = fetch_jobs()
+
+    st.subheader("Best Matches")
+    matches = match_jobs(resume_text, jobs, top_k=5)
+
+    for job, score in matches:
+        st.markdown(f"**{job['position']}** at *{job['company']}*  \n"
+                    f"[View Job Posting]({job['url']})  \n"
+                    f"**Match Score:** {score:.2f}")
+
+        if st.button(f"Generate Resume for {job['position']}", key=job['id']):
+            tailored_resume = generate_resume(resume_text, job)
+            st.text_area("Tailored Resume", tailored_resume, height=300)