- Changed progress callback from "execution_complete" to "workflow_graph_finished" to provide intermediate updates.
- Updated documentation to clarify the distinction between "workflow_graph_finished" and "execution_complete".
- Adjusted frontend API handling to accommodate new event structure and ensure proper result serialization.
Root cause: FastAPI matches routes in definition order. The /{prompt_id}
catch-all at line 257 was intercepting /execute-stream requests before
the specific route handler could match.
Fix: Moved execute-stream definition (with section header + imports)
to line 257, before the catch-all route (now at line 414).
This resolves the 'Connection to server lost' error in SSE streaming.
- /execute-stream now at line 260 (was 1448)
- /{prompt_id} now at line 410 (was 257)
- FastAPI will now match /execute-stream correctly
- Fixes 'Connection to server lost' error in analysis page
ROOT CAUSE FOUND:
FastAPI matches routes in ORDER. The catch-all route /{prompt_id} at line 257
matches /execute-stream BEFORE the specific route at line 1448 can match.
Result: /api/prompts/execute-stream gets routed to get_prompt() which tries
to parse 'execute-stream' as a UUID, causing the error we've been seeing.
SOLUTION: Move /execute-stream route definition to BEFORE line 257 (before /{prompt_id})
This explains why require_auth_flexible was never called - the wrong endpoint
was being invoked entirely.
Root Cause:
- FastAPI cannot use same parameter name in endpoint and dependency
- Query param 'token' conflicted between endpoint and require_auth_flexible
- FastAPI cached dependency signatures at startup
Solution:
- Renamed to 'ssetoken' in require_auth_flexible (backend/auth.py)
- Updated frontend to use ssetoken (frontend/src/utils/api.js)
- Removed debug logging
- Added test endpoint /test-ssetoken
Testing:
✅ Header auth: X-Auth-Token works
✅ Query auth: ?ssetoken=XXX works
✅ SSE streaming: Ready for testing
Note: Required full rebuild, not just restart
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Backend:
- prompts.py: @router.post → @router.get für /execute-stream
- EventSource unterstützt nur GET-Requests
- modules/timeframes nutzen Defaults (SSE kann keine komplexen Params)
Fixes:
- "Connection to server lost" bei Analyse-Ausführung
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
- workflow_executor.py: Generate node_label from prompt_slug or node.type (WorkflowNode has no label attribute)
- prompts.py: Fix INSERT statement - use 'created' column instead of 'created_at'
SSE endpoint now works correctly for workflow execution streaming.
Backend:
- workflow_executor.py: Add progress_callback parameter, emit events for execution_started, node_complete, execution_complete, execution_failed
- prompt_executor.py: Thread progress_callback through execute chain
- routers/prompts.py: New /execute-stream endpoint with asyncio Queue for SSE
Frontend:
- utils/api.js: New executeUnifiedPromptStream() function with EventSource
- pages/Analysis.jsx: Use SSE with live progress display (X/Y Nodes)
Fixes:
- No more gateway timeouts for complex workflows (10+ nodes)
- Live progress feedback for users
- Unlimited workflow complexity
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Bug: debug=true in URL was ignored because FastAPI expected it in
request body (POST without Query() expects body params by default).
Result: node_states were never returned, even with ?debug=true
Fix: Changed debug and save to Query parameters:
- debug: bool = Query(False, ...)
- save: bool = Query(False, ...)
Now ?debug=true in URL correctly enables debug output with node_states.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Problem: Workflows executed via /api/prompts/execute (not /api/workflows/execute)
were passing call_openrouter directly to execute_prompt_with_data, which then
passes it to workflow_executor. workflow_executor expects (prompt, model) signature
but call_openrouter has (prompt, max_tokens=4096) signature.
Previous fix in workflows.py was correct but unused - workflows use prompts.py endpoint.
Solution: Added workflow_llm_call() wrapper in execute_unified_prompt() endpoint
that matches expected (prompt, model) -> str signature.
Related: cb3aa48 (workflows.py fix for different endpoint)
- Increased the maximum token limit in the `call_openrouter` function from 1500 to 4096 to allow for more extensive responses.
- Implemented robust error handling for API requests, including timeout and request errors, with detailed HTTP exceptions for better debugging.
- Improved JSON response handling to ensure valid data is returned, with specific error messages for missing content in the response.
- Enhanced the overall reliability of the OpenRouter API integration, providing clearer feedback for users in case of issues.
These changes improve the user experience by ensuring more comprehensive responses and clearer error reporting during API interactions.
- Updated `resolve_placeholders` in `prompt_executor.py` to support combined modifiers for placeholders, allowing for more flexible output formats.
- Enhanced `build_ai_placeholder_caption` in `placeholder_registry.py` to clarify the generation of AI context captions, focusing on descriptions and explanations.
- Introduced new helper functions in `placeholder_resolver.py` to streamline the retrieval of descriptions and explanations for placeholders.
- Modified tests to cover new functionality, ensuring accurate behavior for combined modifiers and improved placeholder resolution.
These changes enhance the usability and clarity of placeholder outputs, providing users with richer contextual information.
- Updated `build_ai_placeholder_caption` in `placeholder_registry.py` to improve the generation of AI context captions by prioritizing descriptions and avoiding redundancy.
- Introduced `format_value_with_d_modifier` in `placeholder_resolver.py` to format values with contextual information, enhancing the clarity of exported placeholder values.
- Modified `export_placeholder_values` in `prompts.py` to utilize the new formatting function, ensuring that exported data includes both raw values and contextual descriptions.
- Added tests for the new formatting function and updated existing tests to ensure accurate caption generation.
These changes improve the contextual relevance of placeholder data and enhance the user experience when interacting with exported values.
- Introduced `build_ai_placeholder_caption` function in `placeholder_registry.py` to generate AI context captions based on placeholder metadata.
- Updated `resolve_placeholders` in `placeholder_resolver.py` to support modifiers for AI context, allowing for enhanced descriptions when placeholders are resolved.
- Modified `get_placeholder_catalog` to include AI captions in the output, improving the metadata available for placeholders.
- Adjusted `export_placeholder_values` to include AI captions in the exported data, enhancing the information provided to users.
These changes improve the flexibility and functionality of the placeholder system, enabling richer context generation for dynamic content.
- Updated `extract_value_raw` to improve JSON parsing and handle unavailable data more effectively.
- Introduced new functions in `placeholder_resolver.py` for standardized responses when data is unavailable, enhancing clarity for users and AI.
- Modified various data retrieval functions to utilize the new response format, providing detailed reasons for unavailability.
- Improved availability checks in `export_placeholder_values_extended` to account for new response formats.
These changes enhance the robustness of the placeholder system and improve user experience by providing clearer error messages and data handling.
**Error:**
```
psycopg2.ProgrammingError: can't adapt type 'dict'
```
**Root Cause:**
- duplicate_prompt passed Python dicts directly to SQL INSERT
- JSONB fields from r2d() are already deserialized by psycopg2
- PostgreSQL expects JSON strings for JSONB columns
**Fix:**
- Added json.dumps() for all JSONB fields before INSERT:
- stages, output_schema, question_augmentations, graph_data
- Same pattern as import function
Files changed:
- backend/routers/prompts.py: JSON-encode JSONB in duplicate_prompt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
**Root Cause:**
- FastAPI route matching: /{prompt_id} caught ALL requests including /export-all
- Specific routes MUST be defined BEFORE path parameter routes
**Error:**
```
psycopg2.errors.InvalidTextRepresentation: invalid input syntax for type uuid: "export-all"
LINE 1: SELECT * FROM ai_prompts WHERE id='export-all'
```
**Fix:**
- Moved /export-all and /import endpoints to line 106 (BEFORE /{prompt_id} at ~260)
- Added warning comments to both functions
- Fixed typo: for r in → for row in
**Affected:**
- /export-all: Internal Server Error → now works ✅
- /import: Would have had same issue → preemptively fixed ✅
Files changed:
- backend/routers/prompts.py: Reordered route definitions
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Introduced a new utility function to streamline the extraction of user-facing content from aggregated workflow results.
- Updated backend prompt handling to utilize the new function for improved clarity and maintainability.
- Adjusted frontend analysis component to leverage the utility for consistent content display across different workflow result formats.
These changes enhance the overall user experience by ensuring more reliable and readable output from workflow executions.
- Added support for handling aggregated results in workflow prompts, allowing for various data formats (string, object).
- Introduced a utility function to filter active prompts for both pipeline and workflow types in the analysis page.
- Updated content handling in the analysis component to accommodate new workflow data structures.
This improves the flexibility and usability of the prompt execution process in both backend and frontend components.
Backend:
- Moved /placeholders endpoint BEFORE /{prompt_id} catch-all
- Prevents "placeholders" being parsed as UUID parameter
- Fixes 500 Internal Server Error preventing placeholder loading
Frontend:
- PlaceholderPicker can now load ~120+ system placeholders
Root Cause:
- FastAPI matches routes in order
- Generic /{prompt_id} was catching /placeholders first
- psycopg2 error: invalid input syntax for type uuid: "placeholders"
Version: 0.9p (workflow module)
Part 3: End Node Template Engine
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Critical Backend Bug:
- Frontend calls api.getPrompt(id) → GET /api/prompts/{uuid}
- Backend had NO endpoint for single prompt retrieval by ID
- Result: 405 Method Not Allowed
Backend Endpoints Before:
✓ GET /api/prompts - List all
✓ POST /api/prompts - Create
✓ PUT /api/prompts/{id} - Update
✗ GET /api/prompts/{id} - MISSING!
Backend Endpoints After:
✓ GET /api/prompts - List all
✓ GET /api/prompts/{id} - Get single (NEW)
✓ POST /api/prompts - Create
✓ PUT /api/prompts/{id} - Update
Implementation:
- Added get_prompt(prompt_id: str) function
- Returns single prompt by UUID
- 404 if not found
- Requires auth (admin or user)
This fixes:
- Workflow loading after save (loadWorkflow calls getPrompt)
- Workflow editing from admin list (Edit button calls getPrompt)
- All 405 Method Not Allowed errors
Root Cause: Backend was incomplete, missing basic CRUD read-by-id endpoint
Error: NameError TimeWindow not defined
Fix: Graceful degradation if old metadata enums not available
Gap report now optional (empty if old system unavailable)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
NameError: name 'Header' is not defined
Added Header to fastapi imports for export endpoints auth fix.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Backend:
- Add ALL stage outputs to metadata (not just referenced ones)
- Format JSON with indent for readability
- Description: 'Zwischenergebnis aus Stage X'
Frontend:
- Stage raw values shown in collapsible <details> element
- JSON formatted in <pre> tag with syntax highlighting
- 'JSON anzeigen ▼' summary for better UX
Fixes: Stage X - Rohdaten now shows intermediate results
BUG: Wertetabelle wurde nicht angezeigt
FIX: enable_debug=true wenn save=true (für metadata collection)
- metadata wird nur gespeichert wenn debug aktiv
- jetzt: debug or save → metadata immer verfügbar
BUG: {{placeholder|d}} Modifier funktionierte nicht
ROOT CAUSE: catalog wurde bei Exception nicht zu variables hinzugefügt
FIX:
- variables['_catalog'] = catalog (auch wenn None)
- Warning-Log wenn catalog nicht geladen werden kann
- Debug warning wenn |d ohne catalog verwendet
BUG: Platzhalter in Pipeline-Stages am Ende statt an Cursor
FIX:
- stageTemplateRefs Map für alle Stage-Textareas
- onClick + onKeyUp tracking für Cursor-Position
- Insert at cursor: template.slice(0, pos) + placeholder + template.slice(pos)
- Focus + Cursor restore nach Insert
TECHNICAL:
- prompt_executor.py: Besseres Exception Handling für catalog
- UnifiedPromptModal.jsx: Refs für alle Template-Felder
- prompts.py: enable_debug=debug or save
version: 9.6.1 (bugfix)
module: prompts 2.1.1
BREAKING: Analysis page switched from old /insights/run to new /prompts/execute
Changes:
- Backend: Added save=true parameter to /prompts/execute
- When enabled, saves final output to ai_insights table
- Extracts content from pipeline output (last stage)
- Frontend api.js: Added save parameter to executeUnifiedPrompt()
- Frontend Analysis.jsx: Switched from api.runInsight() to api.executeUnifiedPrompt()
- Transforms new result format to match InsightCard expectations
- Pipeline outputs properly extracted and displayed
Fixes: PIPELINE_MASTER responses (old template being sent to AI)
The old /insights/run endpoint used raw template field, which for the
legacy "pipeline" prompt was literally "PIPELINE_MASTER". The new
executor properly handles stages and data processing.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Backend: debug mode in prompt_executor with placeholder tracking
- Backend: show resolved/unresolved placeholders, final prompts, AI responses
- Frontend: test button in UnifiedPromptModal for saved prompts
- Frontend: debug output viewer with JSON preview
- Frontend: wider placeholder example fields in PlaceholderPicker
Resolves pipeline execution debugging issues.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
