feat: Workflow Engine Part 3 - Inline Prompts (v0.9q)
Deploy Development / deploy (push) Successful in 56s
Build Test / pytest-backend (push) Failing after 4s
Build Test / lint-backend (push) Successful in 0s
Build Test / build-frontend (push) Successful in 17s

Ermöglicht Analysis Nodes zwischen zwei Prompt-Modi zu wählen:
- Reference Mode: Basis-Prompt aus DB referenzieren (bestehend)
- Inline Mode: Template direkt im Node editieren (NEU)

Frontend:
- InlineTemplateEditor Component (~80 Zeilen)
- Radio Buttons in WorkflowEditorPage für Mode-Auswahl
- Placeholder Picker für beide Modi (End Node + Inline Template)
- Cursor-Position Tracking mit textareaRef
- Conditional Rendering basierend auf promptSource
- Validation: Entweder prompt_slug ODER inline_template

Backend:
- load_prompt_template() akzeptiert ganzen WorkflowNode (statt nur slug)
- Unterstützt inline_template (Mode 1) und prompt_slug (Mode 2)
- WorkflowNode.inline_template Feld hinzugefügt
- Validation: HTTPException wenn weder slug noch template

Serialization:
- inline_template in graph_data speichern/laden
- Backward-compatible mit bestehenden Workflows

Version: 0.9q
Module: workflow 0.7.0

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
LarsandClaude Opus 4.6 committed 2026-04-11 08:45:00 +02:00
1 parent b453ce63c6
commit a1723db387
7 files changed
+295 -79

No files matched your search

+43 -17
View File
@@ -278,9 +278,10 @@ async def execute_node(
# Analysis Nodes
if node.type == "analysis":
# 1. Lade Prompt
prompt_template = await load_prompt_template(node.prompt_slug, context)
logger.debug(f"Node {node.id}: Loaded prompt '{node.prompt_slug}'")
# 1. Lade Prompt (Part 3: inline_template support)
prompt_template = await load_prompt_template(node, context)
source_type = "inline" if node.inline_template else "reference"
logger.debug(f"Node {node.id}: Loaded prompt from {source_type}")
# 2. Parse question_augmentations
questions = []
@@ -812,39 +813,64 @@ def _has_active_incoming_edge(node, graph: WorkflowGraph, context: Dict[str, Any
return False
async def load_prompt_template(prompt_slug: str, context: Dict[str, Any]) -> str:
async def load_prompt_template(node: WorkflowNode, context: Dict[str, Any]) -> str:
"""
Lädt Prompt-Template aus DB und resolved Platzhalter.
Lädt Prompt-Template aus DB (reference mode) oder direkt vom Node (inline mode).
Part 3: Inline Prompts - Unterstützt zwei Modi:
- Reference Mode: prompt_slug → Template aus ai_prompts Tabelle
- Inline Mode: inline_template → Template direkt vom Node
Args:
prompt_slug: Slug des Prompts (z.B. "pipeline_body")
node: WorkflowNode mit prompt_slug ODER inline_template
context: {"variables": {"name": "Lars", ...}, "profile_id": "..."}
Returns:
Resolved prompt template
Raises:
HTTPException: Wenn weder prompt_slug noch inline_template gesetzt
Beispiel:
>>> template = await load_prompt_template("pipeline_body", {"profile_id": "123"})
>>> node = WorkflowNode(id="n1", prompt_slug="pipeline_body")
>>> template = await load_prompt_template(node, {"profile_id": "123"})
>>> "{{name}}" not in template
True
"""
from placeholder_resolver import get_placeholder_example_values, get_placeholder_catalog
from prompt_executor import resolve_placeholders
from fastapi import HTTPException
with get_db() as conn:
cur = get_cursor(conn)
cur.execute(
"SELECT template FROM ai_prompts WHERE slug = %s AND active = true",
(prompt_slug,)
# Mode 1: Inline Template (NEU)
if node.inline_template:
logger.debug(f"Node {node.id}: Using inline template ({len(node.inline_template)} chars)")
template = node.inline_template
# Mode 2: Reference (bestehend)
elif node.prompt_slug:
logger.debug(f"Node {node.id}: Loading prompt '{node.prompt_slug}' from DB")
with get_db() as conn:
cur = get_cursor(conn)
cur.execute(
"SELECT template FROM ai_prompts WHERE slug = %s AND active = true",
(node.prompt_slug,)
)
row = cur.fetchone()
if not row:
raise HTTPException(status_code=404, detail=f"Prompt not found: {node.prompt_slug}")
template = row['template']
# Mode 3: Error - weder inline noch reference
else:
raise HTTPException(
status_code=400,
detail=f"Node {node.id}: Either prompt_slug or inline_template required"
)
row = cur.fetchone()
if not row:
raise ValueError(f"Prompt not found: {prompt_slug}")
template = row['template']
# Resolve Placeholders using modern prompt_executor method
profile_id = context.get("profile_id")
if not profile_id:
raise HTTPException(status_code=400, detail="profile_id required in context")
# Build variables dict with ALL registered placeholders
variables = {}