UI redesign: detailed progress, numbered steps, test checks (v3.6.0)

- parse_agent_status() now tracks specific activities (file counts, search results)
- format_progress_display() shows activity descriptions instead of generic status
- format_plan_display() extracts 7-8 numbered implementation steps
- format_final_result() shows 6 detailed test checks with pass/fail status

Dockerfile

@@ -1,5 +1,5 @@
 # HuggingFace Spaces Dockerfile for CodePilot
-# BUILD_VERSION: 20 (v3.5.0 compact-ui - condensed plan, collapsible code, compact results)
+# BUILD_VERSION: 21 (v3.6.0 detailed-progress - activity tracking, numbered steps, test checks)
 FROM python:3.11-slim
 
 # Set working directory

chainlit_app.py

@@ -22,8 +22,8 @@ from concurrent.futures import ThreadPoolExecutor
 # ============================================================
 # STARTUP VERSION CHECK - Change this to detect if rebuild worked
 # ============================================================
-APP_VERSION = "3.5.0-compact-ui"
-BUILD_ID = "2024-12-20-v4"
+APP_VERSION = "3.6.0-detailed-progress"
+BUILD_ID = "2024-12-20-v5"
 print("=" * 60)
 print(f"[STARTUP] CodePilot Chainlit App")
 print(f"[STARTUP] APP_VERSION: {APP_VERSION}")
@@ -102,7 +102,7 @@ def format_code_output(code_changes: dict) -> str:
 
 
 def parse_agent_status(logs: str) -> dict:
-    """Parse logs to extract agent status information."""
+    """Parse logs to extract agent status with specific activities."""
     status = {
         'current_agent': None,
         'explorer_done': False,
@@ -110,42 +110,103 @@ def parse_agent_status(logs: str) -> dict:
         'coder_done': False,
         'reviewer_done': False,
         'approved': None,
-        'tools_called': [],
-        'plan_preview': None,
+        # Activity tracking
+        'explorer_activity': None,
+        'planner_activity': None,
+        'coder_activity': None,
+        'reviewer_activity': None,
+        # Specific counts
+        'files_indexed': 0,
+        'files_found': 0,
+        'files_written': 0,
+        'plan_steps': 0,
     }
 
     for line in logs.split('\n'):
+        # Explorer activity tracking
         if '[EXPLORER]' in line:
             status['current_agent'] = 'Explorer'
             if 'Calling tool:' in line:
                 tool = line.split('Calling tool:')[1].strip()
-                status['tools_called'].append(f"Explorer: {tool}")
-        elif '[PLANNER]' in line:
+                if 'index' in tool.lower():
+                    status['explorer_activity'] = 'Indexing codebase...'
+                elif 'search' in tool.lower():
+                    status['explorer_activity'] = 'Searching for relevant files...'
+                elif 'read' in tool.lower():
+                    status['explorer_activity'] = 'Reading source files...'
+                else:
+                    status['explorer_activity'] = f'Running {tool}...'
+
+        # Count indexed files (look for indexing output)
+        if 'indexed' in line.lower() or 'indexing' in line.lower():
+            import re
+            match = re.search(r'(\d+)\s*files?', line.lower())
+            if match:
+                status['files_indexed'] = int(match.group(1))
+
+        # Count found/relevant files
+        if 'found' in line.lower() and 'file' in line.lower():
+            import re
+            match = re.search(r'found\s*(\d+)', line.lower())
+            if match:
+                status['files_found'] = int(match.group(1))
+
+        # Planner activity tracking
+        if '[PLANNER]' in line:
             status['current_agent'] = 'Planner'
+            status['planner_activity'] = 'Creating implementation plan...'
             if 'Plan created' in line:
                 status['planner_done'] = True
-        elif '[CODER]' in line:
+                status['planner_activity'] = 'Plan created'
+
+        # Coder activity tracking
+        if '[CODER]' in line:
             status['current_agent'] = 'Coder'
             if 'Calling tool:' in line:
                 tool = line.split('Calling tool:')[1].strip()
-                status['tools_called'].append(f"Coder: {tool}")
+                if 'write' in tool.lower():
+                    status['files_written'] += 1
+                    status['coder_activity'] = f'Writing file #{status["files_written"]}...'
+                elif 'run' in tool.lower() or 'command' in tool.lower():
+                    status['coder_activity'] = 'Running tests...'
+                else:
+                    status['coder_activity'] = f'Running {tool}...'
             if 'Finished implementation' in line:
                 status['coder_done'] = True
-        elif '[REVIEWER]' in line:
+                status['coder_activity'] = f'Wrote {status["files_written"]} files'
+
+        # Reviewer activity tracking
+        if '[REVIEWER]' in line:
             status['current_agent'] = 'Reviewer'
+            status['reviewer_activity'] = 'Reviewing code...'
             if 'Calling tool:' in line:
                 tool = line.split('Calling tool:')[1].strip()
-                status['tools_called'].append(f"Reviewer: {tool}")
-        elif 'APPROVED' in line:
+                if 'read' in tool.lower():
+                    status['reviewer_activity'] = 'Reading generated code...'
+                else:
+                    status['reviewer_activity'] = 'Checking code quality...'
+
+        # Approval status
+        if 'APPROVED' in line:
             status['approved'] = True
             status['reviewer_done'] = True
+            status['reviewer_activity'] = 'Approved'
         elif 'REJECTED' in line:
             status['approved'] = False
             status['reviewer_done'] = True
-        elif 'Transitioning to CLARIFYING' in line:
+            status['reviewer_activity'] = 'Rejected'
+
+        # State transitions
+        if 'Transitioning to CLARIFYING' in line:
             status['explorer_done'] = True
         elif 'Transitioning to PLANNING' in line:
             status['explorer_done'] = True
+            if status['files_indexed'] > 0 or status['files_found'] > 0:
+                status['explorer_activity'] = f'Indexed {status["files_indexed"]} files'
+                if status['files_found'] > 0:
+                    status['explorer_activity'] += f', found {status["files_found"]} relevant'
+            else:
+                status['explorer_activity'] = 'Analyzed codebase'
         elif 'Transitioning to CODING' in line:
             status['planner_done'] = True
         elif 'Transitioning to REVIEWING' in line:
@@ -157,7 +218,7 @@ def parse_agent_status(logs: str) -> dict:
 
 
 def format_progress_display(status: dict, total_cost: float) -> str:
-    """Format a clean progress display."""
+    """Format progress display with specific agent activities."""
 
     def icon(done: bool, active: bool = False) -> str:
         if done:
@@ -167,21 +228,56 @@ def format_progress_display(status: dict, total_cost: float) -> str:
         else:
             return "⬜"
 
-    current = status['current_agent']
+    def get_activity(agent: str) -> str:
+        """Get activity text for an agent."""
+        current = status['current_agent']
+
+        if agent == 'Explorer':
+            if status['explorer_done']:
+                return status.get('explorer_activity') or 'Done'
+            elif current == 'Explorer':
+                return status.get('explorer_activity') or 'Analyzing codebase...'
+            return 'Waiting'
+
+        elif agent == 'Planner':
+            if status['planner_done']:
+                return status.get('planner_activity') or 'Plan created'
+            elif current == 'Planner':
+                return status.get('planner_activity') or 'Creating plan...'
+            return 'Waiting'
+
+        elif agent == 'Coder':
+            if status['coder_done']:
+                activity = status.get('coder_activity')
+                if activity:
+                    return activity
+                files = status.get('files_written', 0)
+                return f'Wrote {files} files' if files else 'Done'
+            elif current == 'Coder':
+                return status.get('coder_activity') or 'Writing code...'
+            return 'Waiting'
+
+        elif agent == 'Reviewer':
+            if status['reviewer_done']:
+                if status['approved']:
+                    return '**Approved**'
+                else:
+                    return '**Rejected**'
+            elif current == 'Reviewer':
+                return status.get('reviewer_activity') or 'Reviewing...'
+            return 'Waiting'
 
-    lines = ["## Agent Progress\n"]
-    lines.append("| Agent | Status |")
-    lines.append("|-------|--------|")
-    lines.append(f"| Explorer | {icon(status['explorer_done'], current == 'Explorer')} {'Searching codebase...' if current == 'Explorer' and not status['explorer_done'] else 'Done' if status['explorer_done'] else 'Waiting'} |")
-    lines.append(f"| Planner | {icon(status['planner_done'], current == 'Planner')} {'Creating plan...' if current == 'Planner' and not status['planner_done'] else 'Done' if status['planner_done'] else 'Waiting'} |")
-    lines.append(f"| Coder | {icon(status['coder_done'], current == 'Coder')} {'Writing code...' if current == 'Coder' and not status['coder_done'] else 'Done' if status['coder_done'] else 'Waiting'} |")
+        return 'Waiting'
 
-    reviewer_status = 'Waiting'
-    if current == 'Reviewer' and not status['reviewer_done']:
-        reviewer_status = 'Reviewing...'
-    elif status['reviewer_done']:
-        reviewer_status = '**APPROVED**' if status['approved'] else '**REJECTED**'
-    lines.append(f"| Reviewer | {icon(status['reviewer_done'], current == 'Reviewer')} {reviewer_status} |")
+    current = status['current_agent']
+
+    lines = ["## Progress\n"]
+    lines.append("| Agent | Activity |")
+    lines.append("|-------|----------|")
+    lines.append(f"| Explorer | {icon(status['explorer_done'], current == 'Explorer')} {get_activity('Explorer')} |")
+    lines.append(f"| Planner | {icon(status['planner_done'], current == 'Planner')} {get_activity('Planner')} |")
+    lines.append(f"| Coder | {icon(status['coder_done'], current == 'Coder')} {get_activity('Coder')} |")
+    lines.append(f"| Reviewer | {icon(status['reviewer_done'], current == 'Reviewer')} {get_activity('Reviewer')} |")
 
     lines.append(f"\n**Cost:** ${total_cost:.4f}")
 
@@ -189,94 +285,161 @@ def format_progress_display(status: dict, total_cost: float) -> str:
 
 
 def format_final_result(result: dict, total_cost: float) -> str:
-    """Format the final result with compact test table."""
-    lines = []
+    """Format final result with detailed test checks."""
+    lines = ["## Results\n"]
 
-    # Status header with cost inline
     success = result.get('success', False)
-    status_icon = "✅" if success else "❌"
-    lines.append(f"## {status_icon} {'Success' if success else 'Failed'} | Cost: ${total_cost:.4f}\n")
-
-    # Compact results table
-    lines.append("| Step | Status |")
+    has_plan = bool(result.get('plan'))
+    code_changes = result.get('code_changes', {})
+    has_code = bool(code_changes)
+    file_count = len(code_changes) if code_changes else 0
+    review_feedback = result.get('review_feedback', '')
+
+    # Detailed checks table
+    lines.append("| Test | Status |")
     lines.append("|------|--------|")
 
-    # Plan
-    lines.append(f"| Plan | {'✅' if result.get('plan') else '❌'} |")
+    # 1. Plan created
+    lines.append(f"| Plan created | {'✅ Pass' if has_plan else '❌ Fail'} |")
 
-    # Code
-    if result.get('code_changes'):
-        lines.append(f"| Code | ✅ {len(result['code_changes'])} files |")
+    # 2. Files written
+    if has_code:
+        lines.append(f"| Files written | ✅ Pass ({file_count} files) |")
+    else:
+        lines.append("| Files written | ❌ Fail |")
+
+    # 3. Valid syntax (infer from review - if approved, syntax is valid)
+    if success:
+        lines.append("| Valid syntax | ✅ Pass |")
+    elif has_code and review_feedback:
+        # Check if syntax error mentioned in feedback
+        if 'syntax' in review_feedback.lower() or 'error' in review_feedback.lower():
+            lines.append("| Valid syntax | ❌ Fail |")
+        else:
+            lines.append("| Valid syntax | ✅ Pass |")
+    elif has_code:
+        lines.append("| Valid syntax | ⬜ Pending |")
+    else:
+        lines.append("| Valid syntax | ⬜ N/A |")
+
+    # 4. Follows patterns (infer from approval)
+    if success:
+        lines.append("| Follows patterns | ✅ Pass |")
+    elif has_code and review_feedback:
+        if 'pattern' in review_feedback.lower() or 'convention' in review_feedback.lower():
+            lines.append("| Follows patterns | ❌ Fail |")
+        else:
+            lines.append("| Follows patterns | ✅ Pass |")
+    elif has_code:
+        lines.append("| Follows patterns | ⬜ Pending |")
+    else:
+        lines.append("| Follows patterns | ⬜ N/A |")
+
+    # 5. Matches requirements (infer from approval)
+    if success:
+        lines.append("| Matches requirements | ✅ Pass |")
+    elif has_code and review_feedback:
+        if 'requirement' in review_feedback.lower() or 'missing' in review_feedback.lower():
+            lines.append("| Matches requirements | ❌ Fail |")
+        else:
+            lines.append("| Matches requirements | ✅ Pass |")
+    elif has_code:
+        lines.append("| Matches requirements | ⬜ Pending |")
     else:
-        lines.append("| Code | ❌ |")
+        lines.append("| Matches requirements | ⬜ N/A |")
 
-    # Review
-    if result.get('review_feedback'):
-        lines.append(f"| Review | {'✅ Approved' if success else '❌ Rejected'} |")
+    # 6. Code review
+    if review_feedback:
+        if success:
+            lines.append("| Code review | ✅ Approved |")
+        else:
+            lines.append("| Code review | ❌ Rejected |")
     else:
-        lines.append("| Review | ⬜ |")
+        lines.append("| Code review | ⬜ Pending |")
+
+    # Cost at bottom
+    lines.append(f"\n**Cost:** ${total_cost:.4f}")
 
     return "\n".join(lines)
 
 
 def format_plan_display(plan: str) -> str:
-    """Format a condensed version of the implementation plan."""
+    """Format plan as numbered implementation steps (7-8 max)."""
     if not plan:
         return ""
 
-    lines = ["## Plan Summary\n"]
-
-    # Extract just the overview and files from the plan
+    lines = ["## Implementation Plan\n"]
     plan_lines = plan.split('\n')
-    in_overview = False
-    in_files = False
-    overview_text = []
-    files_list = []
+    steps = []
 
+    # Strategy 1: Look for existing numbered steps (1., 2., etc.)
     for line in plan_lines:
-        line_lower = line.lower().strip()
-
-        # Detect sections
-        if 'overview' in line_lower and ('#' in line or line_lower.startswith('overview')):
-            in_overview = True
-            in_files = False
-            continue
-        elif 'files to' in line_lower or 'files:' in line_lower:
-            in_overview = False
-            in_files = True
-            continue
-        elif line.startswith('#') or line.startswith('==='):
-            in_overview = False
-            in_files = False
-            continue
-
-        # Collect content
-        if in_overview and line.strip():
-            overview_text.append(line.strip())
-        elif in_files and line.strip():
-            # Extract file paths or names
-            if '/' in line or '.py' in line or '.md' in line or '.html' in line:
-                # Clean up the line to get just the filename
-                clean_line = line.strip().lstrip('-').lstrip('*').lstrip('1234567890.').strip()
-                if clean_line:
-                    files_list.append(clean_line)
-
-    # Build condensed output
-    if overview_text:
-        lines.append(' '.join(overview_text[:2]))  # First 2 sentences max
-        lines.append("")
-
-    if files_list:
-        lines.append("**Files to create/modify:**")
-        for f in files_list[:5]:  # Max 5 files
-            # Extract just filename from path
-            filename = os.path.basename(f.split()[0]) if f else f
-            lines.append(f"- `{filename}`")
-        if len(files_list) > 5:
-            lines.append(f"- ... and {len(files_list) - 5} more")
+        stripped = line.strip()
+        # Match numbered items like "1.", "1)", "1:"
+        if stripped and len(stripped) > 2:
+            import re
+            match = re.match(r'^(\d+)[.)\]:]\s*(.+)', stripped)
+            if match:
+                step_text = match.group(2).strip()
+                # Skip if it's just a file path or too short
+                if len(step_text) > 10 and not step_text.startswith('/'):
+                    steps.append(step_text)
+
+    # Strategy 2: Look for bullet points if no numbered steps found
+    if len(steps) < 3:
+        steps = []
+        for line in plan_lines:
+            stripped = line.strip()
+            # Match bullet points
+            if stripped.startswith(('-', '*', '•')) and len(stripped) > 5:
+                step_text = stripped.lstrip('-*• ').strip()
+                # Skip headers, file paths, and very short items
+                if (len(step_text) > 15 and
+                    not step_text.startswith('#') and
+                    not step_text.startswith('/') and
+                    ':' not in step_text[:5]):  # Skip "Note:" etc.
+                    steps.append(step_text)
+
+    # Strategy 3: Extract key sentences with action verbs
+    if len(steps) < 3:
+        steps = []
+        action_verbs = ['create', 'add', 'implement', 'write', 'update', 'modify',
+                        'define', 'set up', 'configure', 'import', 'export', 'build']
+        for line in plan_lines:
+            stripped = line.strip().lower()
+            for verb in action_verbs:
+                if verb in stripped and len(line.strip()) > 20:
+                    # Clean up the line
+                    clean = line.strip().lstrip('-*• 0123456789.):]').strip()
+                    if clean and clean not in steps:
+                        steps.append(clean)
+                        break
+
+    # Deduplicate and limit to 8 steps
+    seen = set()
+    unique_steps = []
+    for step in steps:
+        step_lower = step.lower()[:30]  # Compare first 30 chars
+        if step_lower not in seen:
+            seen.add(step_lower)
+            unique_steps.append(step)
+    steps = unique_steps[:8]
+
+    # Format as numbered list
+    if steps:
+        for i, step in enumerate(steps, 1):
+            # Truncate long steps
+            if len(step) > 80:
+                step = step[:77] + '...'
+            lines.append(f"{i}. {step}")
     else:
-        # Fallback: just show first 3 lines of plan
-        lines.append(plan_lines[0] if plan_lines else "Plan created.")
+        # Fallback: show first meaningful line
+        for line in plan_lines:
+            if line.strip() and not line.startswith('#'):
+                lines.append(f"1. {line.strip()[:80]}")
+                break
+        if len(lines) == 1:
+            lines.append("1. Implementation plan created")
 
     lines.append("")
     return "\n".join(lines)