Hello-QM/catgo-LRG / server/catgo/workflow
catgo-workflow
VerifiedCreate and manage computational chemistry workflows with CatGo. Supports VASP, CP2K, ORCA, MLP, LAMMPS. Build OER/HER/CO2RR workflows, geometry optimization, frequency analysis, Gibbs energy calculations.
engine/error_handler.py
View on GitHub"""Handle REMOTE_ERROR tasks: smart recovery, retry, or escalate.
Three-tier error recovery:
Tier 1: Custodian (handles VASP errors at runtime -- already integrated)
Tier 2: Rule-based diagnosis (parse error message, apply known fixes, retry)
Tier 3: Escalate to user (mark PAUSED with diagnosis)
"""
from __future__ import annotations
import logging
from typing import Any
from catgo.workflow.db import WorkflowDB
from catgo.workflow.states import TaskState
logger = logging.getLogger(__name__)
def handle_errors(db: WorkflowDB, workflow_id: str, config: dict[str, Any]) -> list[str]:
"""Process all REMOTE_ERROR tasks: smart recovery, retry, or fail.
Returns list of task IDs that were retried (set back to READY).
"""
from catgo.workflow.engine.smart_recovery import diagnose_and_fix, apply_fix
retry_config = config.get("retry", {})
max_retries = retry_config.get("max_retries", 3)
error_tasks = db.get_tasks_by_status(workflow_id, TaskState.REMOTE_ERROR.value)
retried = []
for task in error_tasks:
task_id = task["id"]
error_type = task.get("error_type", "")
retry_count = task.get("retry_count", 0) or 0
# Transient errors (SSH disconnect, network issues) should NOT
# consume retries or escalate to FAILED. Leave them in
# REMOTE_ERROR — the scanner's recovery pass will reset them
# once the connection is restored.
if error_type == "transient":
logger.debug(
"Task %s: skipping transient error (retry_count stays %d)",
task_id, retry_count,
)
continue
if retry_count >= max_retries:
db.update_task(task_id,
status=TaskState.FAILED.value,
error_message=f"Failed after {max_retries} retries: {task.get('error_message', '')}",
)
logger.warning("Task %s: REMOTE_ERROR -> FAILED (max retries %d)", task_id, max_retries)
continue
# Try smart recovery (Tier 2/3)
diagnosis = diagnose_and_fix(db, task, config)
if diagnosis and diagnosis["tier"] == 2 and diagnosis["fixes"]:
# Tier 2: apply known fix and retry
apply_fix(db, task_id, diagnosis["fixes"])
db.update_task(task_id,
status=TaskState.READY.value,
retry_count=retry_count + 1,
error_message=f"Auto-fix applied: {diagnosis['diagnosis']}",
work_dir=None, # force recompute (prior attempt may have stickied a bad path)
)
retried.append(task_id)
logger.info("Task %s: smart recovery applied (%s)", task_id, diagnosis["diagnosis"])
elif diagnosis and diagnosis["tier"] == 3:
# Tier 3: known error, no auto-fix -- escalate to user
db.update_task(task_id,
status=TaskState.PAUSED.value,
error_message=f"Needs manual review: {diagnosis['diagnosis']}. Original: {task.get('error_message', '')}",
)
logger.info("Task %s: escalated to user (%s)", task_id, diagnosis["diagnosis"])
else:
# Unknown error -- simple retry
db.update_task(task_id,
status=TaskState.READY.value,
retry_count=retry_count + 1,
error_message=None,
work_dir=None, # force recompute (prior attempt may have stickied a bad path)
)
retried.append(task_id)
logger.info("Task %s: REMOTE_ERROR -> READY (retry %d/%d)", task_id, retry_count + 1, max_retries)
return retried
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies. | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |
| meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要". | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |