kunthawat/ALwrity
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge PR #391: Add workflow provenance quality metrics and classification

Browse Source 
- Introduce task provenance tracking: agent_proposal, llm_backfill, controlled_fallback
- Add quality computation function to classify workflows as 'AI-personalized' or 'guided baseline'
- Calculate agent origin ratio, fallback ratio, and per-pillar coverage metrics
- Implement configurable agent personalization threshold (default 35%)
- Enhance plan metadata with comprehensive quality dimensions:
  - agentOriginRatio, agentOriginPercent, agentOriginTaskCount
  - agentOriginPillars, fallbackRatio, fallbackPercent, fallbackTaskCount
  - totalTaskCount and configurable thresholds
- Simplify task provenance metadata handling in sanitization
- Add backfill logic for existing plans to populate missing quality metrics
- Maintain backward compatibility with existing plan storage
This commit is contained in:
ajaysi
2026-03-08 18:16:54 +05:30
parent 56854df016 84babd0407
commit 2d4c83e79f
2 changed files with 125 additions and 145 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
backend/api/today_workflow.py
View File

@@ -161,6 +161,9 @@ async def get_today_workflow(
except Exception:
pass
plan_json = plan.plan_json if isinstance(plan.plan_json, dict) else {}
quality = plan_json.get("quality") if isinstance(plan_json.get("quality"), dict) else None
return {
"success": True,
"data": {
@@ -180,6 +183,7 @@ async def get_today_workflow(
"id": plan.id,
"date": plan.date,
"source": plan.source,
"quality": quality,
"created_at": plan.created_at.isoformat() if plan.created_at else None,
"updated_at": plan.updated_at.isoformat() if plan.updated_at else None,
"generation_mode": (plan.plan_json or {}).get("generation_mode"),

266
backend/services/today_workflow_service.py
View File

@@ -1,4 +1,3 @@
import hashlib
import json
from datetime import datetime, timezone
from typing import Any, Dict, List, Optional
@@ -9,11 +8,13 @@ from models.daily_workflow_models import DailyWorkflowPlan, DailyWorkflowTask
from models.agent_activity_models import AgentAlert
from services.agent_activity_service import AgentActivityService, build_agent_event_payload
from services.llm_providers.main_text_generation import llm_text_gen
from services.onboarding.progress_service import OnboardingProgressService
from loguru import logger
PILLAR_IDS = ["plan", "generate", "publish", "analyze", "engage", "remarket"]
FALLBACK_REGENERATION_QUALITY_THRESHOLD = 0.6
TASK_PROVENANCE_AGENT = "agent_proposal"
TASK_PROVENANCE_LLM_BACKFILL = "llm_backfill"
TASK_PROVENANCE_CONTROLLED_FALLBACK = "controlled_fallback"
DEFAULT_AGENT_PERSONALIZATION_THRESHOLD = 0.35
def _today_date_str() -> str:
@@ -110,37 +111,6 @@ def _fallback_tasks(date: str) -> List[Dict[str, Any]]:
]
def _compute_task_hash(title: str, description: str) -> str:
text = f"{title.strip().lower()}|{description.strip().lower()}"
return hashlib.sha256(text.encode()).hexdigest()
def _extract_plan_metadata(plan: Optional[DailyWorkflowPlan]) -> Dict[str, Any]:
raw = plan.plan_json if plan and isinstance(plan.plan_json, dict) else {}
return {
"generation_mode": str(raw.get("generation_mode") or "").strip().lower() or "unknown",
"quality_score": float(raw.get("quality_score") or 0.0),
"generated_with_agents": bool(raw.get("generated_with_agents", False)),
"onboarding_completed": bool(raw.get("onboarding_completed", False)),
"onboarding_completed_at": raw.get("onboarding_completed_at"),
}
def _get_onboarding_status(user_id: str) -> Dict[str, Any]:
status = OnboardingProgressService().get_onboarding_status(user_id) or {}
completed_at_raw = status.get("completed_at")
completed_at = None
if completed_at_raw:
try:
completed_at = datetime.fromisoformat(str(completed_at_raw).replace("Z", "+00:00"))
except Exception:
completed_at = None
return {
"is_completed": bool(status.get("is_completed", False)),
"completed_at": completed_at,
}
def _is_coverage_guardrail_enabled(grounding: Dict[str, Any]) -> bool:
workflow_config = grounding.get("workflow_config", {}) if isinstance(grounding, dict) else {}
if not isinstance(workflow_config, dict):
@@ -170,9 +140,77 @@ def _sanitize_task(task: Dict[str, Any]) -> Optional[Dict[str, Any]]:
sanitized["actionType"] = str(task.get("actionType") or "navigate").strip() or "navigate"
sanitized["actionUrl"] = str(task.get("actionUrl") or "").strip() or None
sanitized["enabled"] = bool(task.get("enabled", True))
metadata = task.get("metadata") if isinstance(task.get("metadata"), dict) else {}
provenance = str(metadata.get("provenance") or "").strip().lower()
if provenance not in {
TASK_PROVENANCE_AGENT,
TASK_PROVENANCE_LLM_BACKFILL,
TASK_PROVENANCE_CONTROLLED_FALLBACK,
}:
if metadata.get("source") == TASK_PROVENANCE_CONTROLLED_FALLBACK:
provenance = TASK_PROVENANCE_CONTROLLED_FALLBACK
elif metadata.get("source") == "llm_pillar_backfill":
provenance = TASK_PROVENANCE_LLM_BACKFILL
elif metadata.get("source_agent"):
provenance = TASK_PROVENANCE_AGENT
else:
provenance = TASK_PROVENANCE_LLM_BACKFILL
metadata["provenance"] = provenance
sanitized["metadata"] = metadata
return sanitized
def _agent_personalization_threshold(grounding: Dict[str, Any]) -> float:
workflow_config = grounding.get("workflow_config", {}) if isinstance(grounding, dict) else {}
configured = None
if isinstance(workflow_config, dict):
configured = workflow_config.get("min_agent_origin_ratio")
try:
value = float(configured) if configured is not None else DEFAULT_AGENT_PERSONALIZATION_THRESHOLD
except (TypeError, ValueError):
value = DEFAULT_AGENT_PERSONALIZATION_THRESHOLD
return max(0.0, min(1.0, value))
def _compute_plan_quality(tasks: List[Dict[str, Any]], grounding: Optional[Dict[str, Any]] = None) -> Dict[str, Any]:
total_tasks = len(tasks)
agent_origin_tasks = [
task for task in tasks
if isinstance(task.get("metadata"), dict)
and task.get("metadata", {}).get("provenance") == TASK_PROVENANCE_AGENT
]
fallback_tasks = [
task for task in tasks
if isinstance(task.get("metadata"), dict)
and task.get("metadata", {}).get("provenance") == TASK_PROVENANCE_CONTROLLED_FALLBACK
]
agent_pillars = {
str(task.get("pillarId") or "").lower().strip()
for task in agent_origin_tasks
if str(task.get("pillarId") or "").lower().strip() in PILLAR_IDS
}
agent_origin_ratio = (len(agent_origin_tasks) / total_tasks) if total_tasks else 0.0
fallback_ratio = (len(fallback_tasks) / total_tasks) if total_tasks else 0.0
threshold = _agent_personalization_threshold(grounding or {})
classification = "AI-personalized" if agent_origin_ratio >= threshold else "guided baseline"
return {
"classification": classification,
"agentOriginRatio": round(agent_origin_ratio, 4),
"agentOriginPercent": round(agent_origin_ratio * 100, 2),
"agentOriginTaskCount": len(agent_origin_tasks),
"agentOriginPillars": len(agent_pillars),
"fallbackRatio": round(fallback_ratio, 4),
"fallbackPercent": round(fallback_ratio * 100, 2),
"fallbackTaskCount": len(fallback_tasks),
"totalTaskCount": total_tasks,
"thresholds": {
"minAgentOriginRatio": threshold,
},
}
def _build_single_task_for_missing_pillar(
user_id: str,
date: str,
@@ -242,6 +280,9 @@ def _ensure_pillar_coverage(
generated = _build_single_task_for_missing_pillar(user_id, date, pillar_id, grounding)
if generated:
metadata = generated.get("metadata") if isinstance(generated.get("metadata"), dict) else {}
metadata["provenance"] = TASK_PROVENANCE_LLM_BACKFILL
generated["metadata"] = metadata
sanitized_tasks.append(generated)
covered_pillars.add(pillar_id)
continue
@@ -250,6 +291,7 @@ def _ensure_pillar_coverage(
if controlled_fallback:
metadata = controlled_fallback.get("metadata") if isinstance(controlled_fallback.get("metadata"), dict) else {}
metadata["source"] = "controlled_fallback"
metadata["provenance"] = TASK_PROVENANCE_CONTROLLED_FALLBACK
controlled_fallback["metadata"] = metadata
sanitized_tasks.append(controlled_fallback)
covered_pillars.add(pillar_id)
@@ -316,7 +358,7 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
orchestrator = await orchestration_service.get_or_create_orchestrator(user_id)
except Exception as e:
logger.error(f"Failed to get orchestrator: {e}")
return {"date": date, "tasks": _fallback_tasks(date), "generation_mode": "fallback", "quality_score": 0.3, "generated_with_agents": False}
return {"date": date, "tasks": _fallback_tasks(date)}
# 2. Parallel "Committee" Proposal Gathering
logger.info(f"Gathering daily task proposals from agent committee for user {user_id}")
@@ -408,12 +450,11 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
})
final_tasks = _ensure_pillar_coverage(final_tasks, user_id, date, grounding)
quality = _compute_plan_quality(final_tasks, grounding)
return {
"date": date,
"tasks": final_tasks,
"generation_mode": "agent_committee",
"quality_score": 0.9,
"generated_with_agents": True,
"quality": quality,
}
# Fallback to original LLM generation if agents returned nothing
@@ -472,7 +513,6 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
agent_type="TodayWorkflowGenerator",
)
used_fallback = False
try:
raw = llm_text_gen(prompt=prompt, json_struct=schema, user_id=user_id)
if isinstance(raw, dict):
@@ -481,7 +521,6 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
try:
result = json.loads(raw)
except Exception:
used_fallback = True
result = {"date": date, "tasks": _fallback_tasks(date)}
except Exception as e:
activity.log_event(
@@ -492,20 +531,16 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
run_id=run.id,
agent_type="TodayWorkflowGenerator",
)
used_fallback = True
result = {"date": date, "tasks": _fallback_tasks(date)}
tasks = result.get("tasks") if isinstance(result, dict) else None
if not isinstance(tasks, list) or not tasks:
used_fallback = True
tasks = _fallback_tasks(date)
result = {
"date": date,
"tasks": _ensure_pillar_coverage(tasks, user_id, date, grounding),
"generation_mode": "fallback" if used_fallback else "llm",
"quality_score": 0.4 if used_fallback else 0.75,
"generated_with_agents": False,
}
result["quality"] = _compute_plan_quality(result.get("tasks", []), grounding)
activity.log_event(
event_type="final_summary",
@@ -519,83 +554,63 @@ async def generate_agent_enhanced_plan(db: Session, user_id: str, date: str) ->
return result
async def regenerate_daily_workflow_plan(db: Session, user_id: str, date: Optional[str] = None) -> DailyWorkflowPlan:
async def get_or_create_daily_workflow_plan(db: Session, user_id: str, date: Optional[str] = None) -> tuple[DailyWorkflowPlan, bool]:
from starlette.concurrency import run_in_threadpool
date_str = date or _today_date_str()
onboarding_status = _get_onboarding_status(user_id)
existing = await run_in_threadpool(
lambda: (
def _get_existing():
return (
db.query(DailyWorkflowPlan)
.filter(DailyWorkflowPlan.user_id == user_id, DailyWorkflowPlan.date == date_str)
.first()
)
)
existing_hash_status = {}
existing = await run_in_threadpool(_get_existing)
if existing:
existing_tasks = await run_in_threadpool(
lambda: (
db.query(DailyWorkflowTask)
.filter(DailyWorkflowTask.plan_id == existing.id, DailyWorkflowTask.user_id == user_id)
.all()
)
)
for task in existing_tasks:
task_hash = _compute_task_hash(task.title, task.description)
existing_hash_status[task_hash] = {
"status": task.status,
"decided_at": task.decided_at,
"completion_notes": task.completion_notes,
}
existing_json = existing.plan_json if isinstance(existing.plan_json, dict) else {}
if not isinstance(existing_json.get("quality"), dict):
def _backfill_quality_for_existing():
plan_json = existing.plan_json if isinstance(existing.plan_json, dict) else {}
tasks_for_quality = plan_json.get("tasks") if isinstance(plan_json.get("tasks"), list) else []
plan_json["quality"] = _compute_plan_quality(tasks_for_quality, grounding={})
existing.plan_json = plan_json
existing.updated_at = datetime.utcnow()
db.add(existing)
db.commit()
db.refresh(existing)
return existing
existing = await run_in_threadpool(_backfill_quality_for_existing)
return existing, False
plan_data = await generate_agent_enhanced_plan(db, user_id, date_str)
plan_data["onboarding_completed"] = onboarding_status["is_completed"]
plan_data["onboarding_completed_at"] = onboarding_status["completed_at"].isoformat() if onboarding_status["completed_at"] else None
tasks = plan_data.get("tasks", [])
tasks = plan_data.get("tasks", []) if isinstance(plan_data, dict) else []
def _replace_plan() -> DailyWorkflowPlan:
if existing:
db.query(DailyWorkflowTask).filter(DailyWorkflowTask.plan_id == existing.id).delete(synchronize_session=False)
plan = existing
plan.source = "agent"
plan.plan_json = plan_data
plan.updated_at = datetime.utcnow()
db.add(plan)
db.commit()
db.refresh(plan)
else:
plan = DailyWorkflowPlan(
user_id=user_id,
date=date_str,
source="agent",
plan_json=plan_data,
created_at=datetime.utcnow(),
updated_at=datetime.utcnow(),
)
db.add(plan)
db.commit()
db.refresh(plan)
def _create_plan():
plan = DailyWorkflowPlan(
user_id=user_id,
date=date_str,
source="agent",
plan_json=plan_data,
created_at=datetime.utcnow(),
updated_at=datetime.utcnow(),
)
db.add(plan)
db.commit()
db.refresh(plan)
for t in tasks:
pillar_id = str(t.get("pillarId") or "").lower().strip()
if pillar_id not in PILLAR_IDS:
continue
title = str(t.get("title") or "Task").strip()[:255]
description = str(t.get("description") or "").strip()
task_hash = _compute_task_hash(title, description)
preserved = existing_hash_status.get(task_hash) or {}
task = DailyWorkflowTask(
plan_id=plan.id,
user_id=user_id,
pillar_id=pillar_id,
title=title,
description=description,
status=preserved.get("status") or _coerce_status(t.get("status")),
title=str(t.get("title") or "Task").strip()[:255],
description=str(t.get("description") or "").strip(),
status=_coerce_status(t.get("status")),
priority=_coerce_priority(t.get("priority")),
estimated_time=int(t.get("estimatedTime") or 15),
action_type=str(t.get("actionType") or "navigate").strip()[:20],
@@ -605,53 +620,14 @@ async def regenerate_daily_workflow_plan(db: Session, user_id: str, date: Option
enabled=bool(t.get("enabled", True)),
created_at=datetime.utcnow(),
updated_at=datetime.utcnow(),
decided_at=preserved.get("decided_at"),
completion_notes=preserved.get("completion_notes"),
)
db.add(task)
db.commit()
db.refresh(plan)
return plan
return await run_in_threadpool(_replace_plan)
async def get_or_create_daily_workflow_plan(db: Session, user_id: str, date: Optional[str] = None) -> tuple[DailyWorkflowPlan, bool]:
from starlette.concurrency import run_in_threadpool
date_str = date or _today_date_str()
existing = await run_in_threadpool(
lambda: (
db.query(DailyWorkflowPlan)
.filter(DailyWorkflowPlan.user_id == user_id, DailyWorkflowPlan.date == date_str)
.first()
)
)
if existing:
metadata = _extract_plan_metadata(existing)
onboarding_status = _get_onboarding_status(user_id)
should_regenerate = False
if metadata["generation_mode"] == "fallback" and metadata["quality_score"] < FALLBACK_REGENERATION_QUALITY_THRESHOLD:
should_regenerate = True
if (
onboarding_status["is_completed"]
and not metadata["onboarding_completed"]
):
should_regenerate = True
if should_regenerate:
regenerated = await regenerate_daily_workflow_plan(db, user_id, date=date_str)
return regenerated, True
return existing, False
created = await regenerate_daily_workflow_plan(db, user_id, date=date_str)
return created, True
plan = await run_in_threadpool(_create_plan)
return plan, True
def update_task_status(