# **Main Code Backup:** import gradio as gr import requests import json import datetime import os OLLAMA_URL = "http://localhost:11434/api/generate" HISTORY_FILE = "interview_history.json" \# ── Persistent History Helpers ───────────────────────────────────────────────── def load_history(): """Load history from JSON file on disk""" if os.path.exists(HISTORY_FILE): with open(HISTORY_FILE, "r") as f: return json.load(f) return \[] def save_history(history): """Save history to JSON file on disk""" with open(HISTORY_FILE, "w") as f: json.dump(history, f, indent=2) \# ── Ollama helper ────────────────────────────────────────────────────────────── def ask_ollama(prompt, model="mistral:7b", temperature=0.7): try: payload = { "model": model, "prompt": prompt, "stream": False, "temperature": temperature } r = requests.post(OLLAMA_URL, json=payload, timeout=300) return r.json().get("response", "No response") except Exception as e: return f"❌ Error: {str(e)}" \# ── Question generation ──────────────────────────────────────────────────────── QUESTION_PROMPTS = \[ "Generate ONE interview question about the candidate's most recent project experience. Question:", "Generate ONE follow-up interview question about the specific technologies or tools used. Question:", "Generate ONE interview question about a challenge they faced and how they overcame it. Question:", ] def clean_question(raw): if "Question:" in raw: raw = raw.split("Question:")\[-1].strip() if "?" in raw: raw = raw.split("?")\[0] + "?" return raw.strip() def generate_all_questions(job_desc, history_state): if not job_desc or len(job_desc.strip()) < 20: return "Please paste a job description (20+ characters).", "", "Ready", history_state, gr.update() job_desc = job_desc\[:500] questions = \[] for p in QUESTION_PROMPTS: full_prompt = f"Based on this job description:\\n{job_desc}\\n\\n{p}" q = clean_question(ask_ollama(full_prompt, temperature=0.8)) questions.append(q) session = { "timestamp": datetime.datetime.now().strftime("%Y-%m-%d %H:%M"), "job_snippet": job_desc\[:60] + "...", "questions": questions, "answers": \["", "", ""], "scores": \["", "", ""], "numeric_scores": \[], } history_state = history_state or \[] history_state.append(session) save_history(history_state) # 💾 Save to disk tips_md = build_tips(job_desc) return questions\[0], "0", "Question 1 / 3", history_state, gr.update(value=tips_md) \# ── Answer scoring ───────────────────────────────────────────────────────────── def score_answer(answer, q_index_str, history_state): if not answer or len(answer.strip()) < 15: return "Please write a longer answer (15+ characters).", history_state idx = int(q_index_str) if q_index_str else 0 answer = answer\[:500] prompt = f"""You are a strict interview coach evaluating a candidate's spoken interview answer. Candidate's answer: {answer} STEP 1 — Relevance check: Is this a genuine attempt at answering an interview question? It is NOT relevant if it is: random text, mix of random text, code, gibberish, a single word, copy-pasted content, or completely off-topic, only partially meaningful or relevant. STEP 2 - If NOT relevant, respond with ONLY these two lines and nothing else: Relevant: NO Score: NIL/10 Warning: ⚠️ Irrelevant response detected. Please answer the interview question properly using the STAR format. If it IS a genuine interview answer, respond with ALL of the following lines and NOTHING else: Relevant: YES Score: X/10 Strength: (one sentence about what was done well) Weakness: (one sentence about the biggest gap) Fix: (one specific, actionable improvement the candidate can make to their spoken answer — no code)""" feedback = ask_ollama(prompt, temperature=0.5) # Save to history if history_state: last = history_state\[-1] if idx < len(last\["answers"]): last\["answers"]\[idx] = answer\[:80] + "..." for line in feedback.splitlines(): if line.startswith("Score:"): score_str = line.replace("Score:", "").strip() last\["scores"]\[idx] = score_str # Extract numeric score try: numeric = float(score_str.split("/")\[0].strip()) last\["numeric_scores"].append(numeric) except: pass break history_state\[-1] = last save_history(history_state) # 💾 Save to disk return feedback, history_state \# ── Navigation ───────────────────────────────────────────────────────────────── def next_question(q_index_str, answer, history_state): """Move to next question, returning previous Q+A for the review panel.""" idx = int(q_index_str) if q_index_str else 0 if not history_state: return "Start an interview first.", "", str(idx), "No session", history_state, "", "", "" session = history_state\[-1] questions = session\["questions"] next_idx = idx + 1 # Capture what the user just answered (for the prev panel) prev_q = questions\[idx] prev_a = answer or "(no answer given)" prev_score = session\["scores"]\[idx] if session\["scores"]\[idx] else "(no feedback yet)" if next_idx >= len(questions): session_log = render_session_log(session, up_to=idx) return ( "✅ All 3 questions complete! Check your history.", "", # clear answer box str(idx), "Interview Complete 🎉", history_state, prev_q, prev_a, session_log, ) session_log = render_session_log(session, up_to=idx) return ( questions\[next_idx], "", # clear answer box str(next_idx), f"Question {next_idx + 1} / 3", history_state, prev_q, prev_a, session_log, ) def render_session_log(session, up_to): """Render completed Q+A+Score pairs for the current session (up to index `up\_to`).""" if up_to < 0: return "No completed questions yet." lines = \[] for i in range(up_to + 1): q = session\["questions"]\[i] a = session\["answers"]\[i] or "(no answer saved)" sc = session\["scores"]\[i] or "(no feedback yet)" lines.append(f"\*\*Q{i+1}:\*\* {q}") lines.append(f"\*Your answer:\* {a}") lines.append(f"\*Score:\* {sc}") lines.append("") return "\\n".join(lines) \# ── History rendering ────────────────────────────────────────────────────────── def compute_stats(history): """Compute overall stats across all sessions""" all_scores = \[] for s in history: for score in s.get("numeric_scores", \[]): all_scores.append(score) if not all_scores: return None avg = sum(all_scores) / len(all_scores) best = max(all_scores) # Trend: compare first half vs second half trend = "" if len(all_scores) >= 4: mid = len(all_scores) // 2 first_avg = sum(all_scores\[:mid]) / mid second_avg = sum(all_scores\[mid:]) / (len(all_scores) - mid) diff = second_avg - first_avg if diff > 0.5: trend = "📈 Improving!" elif diff < -0.5: trend = "📉 Declining — practice more" else: trend = "➡️ Consistent" return { "total_sessions": len(history), "total_answers": len(all_scores), "avg_score": round(avg, 1), "best_score": round(best, 1), "trend": trend } def render_history(history_state): """Render full history with stats""" history = history_state or \[] # Load from disk too (in case state and disk diverge) disk_history = load_history() if len(disk_history) > len(history): history = disk_history if not history: return "No sessions yet. Start your first interview above!" lines = \[] # Stats block stats = compute_stats(history) if stats: lines.append("## 📊 Your Progress") lines.append(f"| Metric | Value |") lines.append(f"|--------|-------|") lines.append(f"| Total Sessions | {stats\['total_sessions']} |") lines.append(f"| Total Answers | {stats\['total_answers']} |") lines.append(f"| Average Score | {stats\['avg_score']}/10 |") lines.append(f"| Best Score | {stats\['best_score']}/10 |") if stats\['trend']: lines.append(f"| Trend | {stats\['trend']} |") lines.append("") # Sessions lines.append("## 📝 Session History") for i, s in enumerate(reversed(history), 1): scores_display = " | ".join(s\["scores"]) if any(s\["scores"]) else "No feedback yet" avg = "" if s.get("numeric_scores"): avg = f" · Avg: {round(sum(s\['numeric_scores'])/len(s\['numeric_scores']), 1)}/10" lines.append(f"### Session {len(history) - i + 1} — {s\['timestamp']}{avg}") lines.append(f"\*\*Role:\*\* {s\['job_snippet']}") lines.append(f"\*\*Scores:\*\* {scores_display}") for j, (q, a, sc) in enumerate(zip(s\["questions"], s\["answers"], s\["scores"]), 1): lines.append(f"\\n### \*\*Q{j}:\*\* {q}\\n") if a: lines.append(f"\*\*Answer:\*\* {a}\\n") if sc: lines.append(f"\*\*Score:\*\* {sc}\\n") lines.append("\\n---") return "\\n".join(lines) \# ── Tips \& Resources ─────────────────────────────────────────────────────────── TIPS_DB = { "python": { "label": "Python / Backend", "leetcode": \[ ("Two Sum", "https://leetcode.com/problems/two-sum/", "Easy"), ("LRU Cache", "https://leetcode.com/problems/lru-cache/", "Medium"), ("Word Search II", "https://leetcode.com/problems/word-search-ii/", "Hard"), ], "concepts": \["OOP principles", "Decorators \& generators", "Async / await", "REST API design"], }, "react": { "label": "React / Frontend", "leetcode": \[ ("Valid Parentheses", "https://leetcode.com/problems/valid-parentheses/", "Easy"), ("Flatten Nested List", "https://leetcode.com/problems/flatten-nested-list-iterator/", "Medium"), ], "concepts": \["Virtual DOM", "Hooks \& state management", "Component lifecycle", "Web performance"], }, "machine learning": { "label": "Machine Learning", "leetcode": \[ ("Find Peak Element", "https://leetcode.com/problems/find-peak-element/", "Medium"), ("Kth Largest Element", "https://leetcode.com/problems/kth-largest-element-in-an-array/", "Medium"), ], "concepts": \["Bias-variance tradeoff", "Overfitting \& regularisation", "Gradient descent", "Model evaluation metrics"], }, "sql": { "label": "SQL / Databases", "leetcode": \[ ("Employees Earning More Than Manager", "https://leetcode.com/problems/employees-earning-more-than-their-managers/", "Easy"), ("Department Top 3 Salaries", "https://leetcode.com/problems/department-top-three-salaries/", "Hard"), ], "concepts": \["JOINs \& subqueries", "Indexing strategies", "Transactions \& ACID", "Query optimisation"], }, } DEFAULT_TIPS = { "label": "General Software Engineering", "leetcode": \[ ("Two Sum", "https://leetcode.com/problems/two-sum/", "Easy"), ("Merge Intervals", "https://leetcode.com/problems/merge-intervals/", "Medium"), ("Trapping Rain Water", "https://leetcode.com/problems/trapping-rain-water/", "Hard"), ], "concepts": \["STAR answer format", "System design basics", "Time \& space complexity", "Behavioural questions"], } def build_tips(job_desc=""): jd_lower = job_desc.lower() matched = DEFAULT_TIPS for key, data in TIPS_DB.items(): if key in jd_lower: matched = data break lc_rows = "\\n".join( f"| \[{p}]({url}) | {diff} |" for p, url, diff in matched\["leetcode"] ) concept_rows = "\\n".join(f"- ✅ {c}" for c in matched\["concepts"]) return f"""## 🎯 Tips for: {matched\['label']} \### 📚 Key Concepts to Revise {concept_rows} \### 💻 Recommended LeetCode Problems | Problem | Difficulty | |---------|-----------| {lc_rows} \### 🧠 General Interview Advice \- Use the \*\*STAR format\*\* (Situation → Task → Action → Result) for behavioural questions \- Always quantify impact: \*"reduced load time by 40%"\* beats \*"made it faster"\* \- It's okay to take 30 seconds to think before answering \- Ask clarifying questions — interviewers reward curiosity \### 🔗 Useful Resources \- \[Grokking the System Design Interview](https://www.educative.io/courses/grokking-the-system-design-interview) \- \[NeetCode Roadmap](https://neetcode.io/roadmap) \- \[Tech Interview Handbook](https://www.techinterviewhandbook.org/) \- \[Pramp — Free Mock Interviews](https://www.pramp.com/) """ \# ── Custom CSS ───────────────────────────────────────────────────────────────── CUSTOM_CSS = """ @import url('https://fonts.googleapis.com/css2?family=Google+Sans:wght@400;500;700\&display=swap'); \* { font-family: 'Google Sans', 'Product Sans', sans-serif !important; } .gr-button-primary { background: linear-gradient(135deg, #6366f1, #8b5cf6) !important; border: none !important; color: white !important; font-weight: 600 !important; transition: transform 0.15s, box-shadow 0.15s !important; } .gr-button-primary:hover { transform: translateY(-2px) !important; box-shadow: 0 6px 20px rgba(99,102,241,0.4) !important; } .gr-button-secondary { border: 2px solid #6366f1 !important; color: #6366f1 !important; font-weight: 600 !important; transition: all 0.15s !important; } .gr-button-secondary:hover { background: #6366f1 !important; color: white !important; } \#progress_box textarea { font-weight: 700 !important; font-size: 1rem !important; color: #6366f1 !important; text-align: center !important; } \#feedback_box textarea { font-size: 0.95rem !important; line-height: 1.6 !important; } .tab-nav button { font-weight: 600 !important; } /\* ── Animated blob background ── \*/ body { background: #0d0d1f !important; } gradio-app { background: transparent !important; } .gradio-container { background: transparent !important; } """ \# ── Build UI ─────────────────────────────────────────────────────────────────── with gr.Blocks( title="AI Interview Coach", theme=gr.themes.Soft( primary_hue=gr.themes.colors.violet, secondary_hue=gr.themes.colors.purple, neutral_hue=gr.themes.colors.slate, font=\[gr.themes.GoogleFont("Google Sans"), "ui-sans-serif", "sans-serif"], ), css=CUSTOM_CSS ) as demo: # Load history from disk on startup history_state = gr.State(load_history()) q_index = gr.State("0") gr.HTML(""" `
`

AI Interview Coach `

` `

` ``Practice · Get Feedback · Improve `
` ``Powered by ``Mistral 7B `

` `
` """) gr.HTML(""" ``
`
`
`
`
`
` """) with gr.Tabs(): # ── Tab 1: Practice ─────────────────────────────────────────────────── with gr.Tab("🎯 Practice"): with gr.Row(): with gr.Column(scale=1): gr.Markdown("### 📋 Step 1 — Paste the Job Description") job_desc_box = gr.Textbox( label="Job Description", lines=6, placeholder="Paste from LinkedIn, company careers page, etc.", ) start_btn = gr.Button("🚀 Start Interview", variant="primary", size="lg") with gr.Column(scale=1): gr.Markdown("### 💬 Step 2 — Answer Questions") progress_box = gr.Textbox( label="Progress", value="Ready to start", interactive=False, elem_id="progress_box", ) question_box = gr.Textbox( label="Interview Question", lines=3, interactive=False, ) answer_box = gr.Textbox( label="Your Answer", lines=5, placeholder="Be specific — use the STAR format (Situation, Task, Action, Result).", ) with gr.Row(): feedback_btn = gr.Button("📊 Get Feedback", variant="primary") next_btn = gr.Button("➡️ Next Question", variant="secondary") # ── Previous answer review panel ────────────────────────────────── with gr.Accordion("🔁 Previous Question Review", open=False) as prev_accordion: with gr.Row(): prev_question_box = gr.Textbox( label="Previous Question", interactive=False, lines=2, placeholder="Will show the last question after you click Next Question...", ) prev_answer_box = gr.Textbox( label="Your Previous Answer", interactive=False, lines=3, placeholder="Will show your last answer here...", ) # ── Full session log accordion ───────────────────────────────────── with gr.Accordion("📋 This Session's Log", open=False): session_log_display = gr.Markdown("Complete questions to see your session log here.") gr.Markdown("### 🏆 Step 3 — Coach Feedback") feedback_box = gr.Textbox( label="AI Coach Feedback", interactive=False, lines=7, elem_id="feedback_box", ) # ── Tab 2: History ──────────────────────────────────────────────────── with gr.Tab("📈 History \& Progress"): gr.Markdown(""" ### Your Interview Sessions History is \*\*saved to disk\*\* — persists across restarts! 💾 """) with gr.Row(): refresh_btn = gr.Button("🔄 Refresh History", variant="secondary") clear_btn = gr.Button("🗑️ Clear History", variant="secondary") history_display = gr.Markdown( render_history(load_history()) # Load from disk on startup ) # ── Tab 3: Tips ─────────────────────────────────────────────────────── with gr.Tab("💡 Tips \& Resources"): tips_display = gr.Markdown(build_tips()) # ── Wire up events ───────────────────────────────────────────────────────── start_btn.click( fn=generate_all_questions, inputs=\[job_desc_box, history_state], outputs=\[question_box, q_index, progress_box, history_state, tips_display], ).then( fn=render_history, inputs=\[history_state], outputs=\[history_display], ) feedback_btn.click( fn=score_answer, inputs=\[answer_box, q_index, history_state], outputs=\[feedback_box, history_state], ).then( fn=render_history, inputs=\[history_state], outputs=\[history_display], ) next_btn.click( fn=next_question, inputs=\[q_index, answer_box, history_state], outputs=\[question_box, answer_box, q_index, progress_box, history_state, prev_question_box, prev_answer_box, session_log_display], ) refresh_btn.click( fn=render_history, inputs=\[history_state], outputs=\[history_display], ) def clear_history(history_state): """Clear all history""" if os.path.exists(HISTORY_FILE): os.remove(HISTORY_FILE) return \[], "History cleared! Start a new interview above." clear_btn.click( fn=clear_history, inputs=\[history_state], outputs=\[history_state, history_display], ) demo.launch(share = True)