issueloop 0.2.0__tar.gz

issueloop-0.2.0/LICENSE

@@ -0,0 +1,23 @@
+LICENSE (MIT)
+
+MIT License
+
+Copyright (c) 2026 amareshhebbar
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

issueloop-0.2.0/PKG-INFO

@@ -0,0 +1,261 @@
+Metadata-Version: 2.4
+Name: issueloop
+Version: 0.2.0
+Summary: Finds failing tests, splits failures into independent tickets, stores them, hands them out one at a time.
+Author: onenot8
+License: LICENSE (MIT)
+        
+        MIT License
+        
+        Copyright (c) 2026 amareshhebbar
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: requests
+Requires-Dist: pyyaml
+Requires-Dist: pathspec
+Provides-Extra: supabase
+Requires-Dist: supabase; extra == "supabase"
+Requires-Dist: python-dotenv; extra == "supabase"
+Provides-Extra: dev
+Requires-Dist: pytest; extra == "dev"
+Dynamic: license-file
+
+# IssueLoop
+
+Finds failing tests (batch or live). Splits failures into independent
+tickets via an LLM, with automatic priority fallback across providers.
+Stores them. Hands them out one at a time. Optionally proposes and
+applies fixes under an explicit, auditable permission allowlist.
+
+![license](https://img.shields.io/badge/license-MIT-blue)
+![python](https://img.shields.io/badge/python-3.10%2B-blue)
+![status](https://img.shields.io/badge/status-phase--2-orange)
+
+---
+
+## Why this exists
+
+Bugs sit unfixed because someone has to notice them first. IssueLoop
+reads the test suite directly — a failing test is the bug report. It
+can also watch a live-running process or log file, for bugs that only
+show up outside a test run. No human has to spot it, and no human has
+to manually triage which failures are actually distinct problems.
+
+## Install
+
+Not yet on PyPI — install from git:
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate
+pip install "git+https://github.com/onenot8/issueLoop@phase-2"
+```
+
+Or for local development:
+
+```bash
+git clone https://github.com/onenot8/issueLoop
+cd issueLoop
+pip install -e ".[dev]"
+```
+
+Local SQLite backend, zero setup. No database account, no API key required to try it.
+
+## Quickstart
+
+```python
+import issueloop
+
+issueloop.use(
+    database="local",                     # or "supabase" for a hosted backup
+    llm={"providers": [                   # priority list — falls back automatically
+        {"provider": "anthropic", "model": "claude-sonnet-4-6", "api_key": "sk-..."},
+        {"provider": "ollama", "model": "qwen2.5-coder:7b"},   # free, local, default if omitted
+    ]},
+    notify={"webhook": "https://your-endpoint"},   # fires if IssueLoop itself breaks
+)
+
+issueloop.scan_repo("repos/myrepo")
+issueloop.run_tests("myrepo")
+issueloop.create_tickets("myrepo")
+
+ticket = issueloop.get_top_error("myrepo")   # claims one, marks in_progress
+issueloop.resolve(ticket["id"])              # your side fixed it
+```
+
+## Live monitoring
+
+Watch a process IssueLoop spawns itself, or tail a log file an
+already-running process writes to:
+
+```python
+from issueloop import watch_process, watch_log_file, stop_watch
+
+handle = watch_process("myrepo", "python3 main.py", cwd="/path/to/repo", debounce_seconds=3.0)
+# ...
+stop_watch(handle)
+```
+
+Or from the CLI, in the foreground:
+
+```bash
+issueloop watch myrepo --command "python3 main.py" --cwd /path/to/repo --debounce 3.0
+issueloop watch myrepo --log-file /path/to/app.log --debounce 3.0
+```
+
+Errors are debounced (default 3s of quiet) so a single multi-line
+traceback becomes one ticket, not dozens. Detected errors land in the
+same log format the batch test runner uses — `issueloop tickets` picks
+them up with no extra steps.
+
+## Fix-apply layer
+
+Off by default — every command needs an explicit allowlist entry:
+
+```yaml
+# config/permission.yaml
+per_repo:
+  myrepo:
+    allowed_patterns:
+      - 'sed -i .* somefile\.py'
+```
+
+```python
+issueloop.propose_fix(ticket_id, "sed -i 's/old/new/' somefile.py")
+result = issueloop.apply_fix(ticket_id)
+# {"status": "resolved" | "retry" | "escalated" | "denied", ...}
+```
+
+`apply_fix` re-runs the ticket's associated test after applying the
+fix and only resolves it if the test actually passes. Failures retry
+up to a configurable limit, then escalate to `needs_human`. Every
+permission decision is audited — `issueloop.get_permission_audit_log()`.
+
+## Use it from the CLI
+
+```bash
+issueloop check-env
+issueloop scan repos/myrepo
+issueloop test myrepo
+issueloop tickets myrepo
+issueloop next myrepo
+issueloop watch myrepo --command "..." | --log-file ...
+issueloop cleanup --days 30
+```
+
+## Use it from Node, Go, or anything else
+
+```bash
+issueloop serve --port 8787
+```
+
+```js
+const r = await fetch("http://localhost:8787/errors/top?repo=myrepo");
+const ticket = await r.json();
+```
+
+`serve` is localhost-only, no auth — put it behind your own gateway if
+you ever expose it further.
+
+## Full API
+
+Bug query — `get_all_bugs`, `get_unresolved_bugs`, `get_resolved_bugs`,
+`get_failed_bugs`, `get_bugs_needing_human`, `get_bugs_by_status`,
+`get_bugs_by_priority`, `get_bug`, `get_bug_count`,
+`get_bug_count_by_status`, `search_bugs`, `get_oldest_bug`, `get_newest_bug`
+
+Lifecycle — `get_top_error`, `get_all_errors`, `resolve`, `fail`,
+`escalate`, `reassign`, `retry_bug`, `get_bug_attempts`, `bulk_resolve`
+
+Fix-apply — `propose_fix`, `apply_fix`, `check_permission`,
+`get_permission_audit_log`
+
+Scan / test — `scan_repo`, `get_file_inventory`, `run_tests`,
+`run_single_test`, `create_tickets`
+
+Live monitoring — `watch_process`, `watch_log_file`, `stop_watch`,
+`list_active_watchers`
+
+LLM / tokens — `get_token_consumption`,
+`get_token_consumption_by_provider`, `get_llm_call_history`,
+`get_llm_provider_status`
+
+Database — `cleanup`, `purge_repo`, `get_database_stats`,
+`export_bugs`, `reap_stale_bugs`, `rotate_logs`
+
+Notifications — `get_crash_log`, `get_notification_config`
+
+Config / utility — `use`, `get_config`, `list_repos`, `health_check`
+
+## Config — every option
+
+| Field                 | Values                                | Default              |
+| ---------------------- | -------------------------------------- | --------------------- |
+| `database`             | `"local"`, `"supabase"`                | `"local"`             |
+| `database_path`        | any path                               | `data/issueloop.db`   |
+| `retention_days`       | int                                    | `30`                  |
+| `llm.providers`        | list of provider dicts, priority order | single ollama default |
+| `llm.provider`         | `"ollama"`, `"anthropic"`, `"openai"`  | `"ollama"`             |
+| `llm.apiKey`           | string                                 | none (required for anthropic/openai) |
+| `llm.tokenSize`        | int                                    | `1024`                |
+| `notify.webhook`       | URL                                    | none                  |
+
+Both `apiKey`/`api_key` and `tokenSize`/`token_size` work — camelCase
+or snake_case, your call. A single `llm={...}` dict still works exactly
+as before; wrap multiple in `llm={"providers": [...]}` for automatic
+fallback.
+
+Config files (`config/permission.yaml`, `config/provider_config.yaml`)
+resolve in this order: an explicit env var
+(`ISSUELOOP_PERMISSION_PATH` / `ISSUELOOP_PROVIDER_CONFIG_PATH`) →
+`./config/` relative to your current directory → `config/` in the
+IssueLoop checkout → bundled package defaults. This means a real
+`pip install` (not just `-e .`) still works correctly even without a
+project-local `config/` directory.
+
+## Why trust this
+
+- Every module has a test, and the tests run against real behavior —
+  a real SQLite file, a real HTTP server on a real socket — not just
+  mocks pretending things work.
+- Real bugs were caught and fixed by actually running this code, not
+  by reading it — including a prompt that only ever sent `stderr`
+  to the LLM triage step while pytest (and most test runners) report
+  failures on `stdout`, and a config-bundling gap that made the
+  permission system silently deny everything under a real (non-editable)
+  install. Both have regression coverage now.
+- Local-first by default. Nothing phones home, nothing requires an
+  account to try.
+- Read the code — it's small enough to actually read. That's on
+  purpose. File-by-file build order: `BUILD_ORDER.md`.
+
+## Testing
+
+```bash
+pytest tests/ -v
+```
+
+Full walkthrough, including live monitoring and the fix-apply layer: `TESTING.md`.
+
+## License
+
+MIT — see `LICENSE`.

issueloop-0.2.0/README.md

@@ -0,0 +1,220 @@
+# IssueLoop
+
+Finds failing tests (batch or live). Splits failures into independent
+tickets via an LLM, with automatic priority fallback across providers.
+Stores them. Hands them out one at a time. Optionally proposes and
+applies fixes under an explicit, auditable permission allowlist.
+
+![license](https://img.shields.io/badge/license-MIT-blue)
+![python](https://img.shields.io/badge/python-3.10%2B-blue)
+![status](https://img.shields.io/badge/status-phase--2-orange)
+
+---
+
+## Why this exists
+
+Bugs sit unfixed because someone has to notice them first. IssueLoop
+reads the test suite directly — a failing test is the bug report. It
+can also watch a live-running process or log file, for bugs that only
+show up outside a test run. No human has to spot it, and no human has
+to manually triage which failures are actually distinct problems.
+
+## Install
+
+Not yet on PyPI — install from git:
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate
+pip install "git+https://github.com/onenot8/issueLoop@phase-2"
+```
+
+Or for local development:
+
+```bash
+git clone https://github.com/onenot8/issueLoop
+cd issueLoop
+pip install -e ".[dev]"
+```
+
+Local SQLite backend, zero setup. No database account, no API key required to try it.
+
+## Quickstart
+
+```python
+import issueloop
+
+issueloop.use(
+    database="local",                     # or "supabase" for a hosted backup
+    llm={"providers": [                   # priority list — falls back automatically
+        {"provider": "anthropic", "model": "claude-sonnet-4-6", "api_key": "sk-..."},
+        {"provider": "ollama", "model": "qwen2.5-coder:7b"},   # free, local, default if omitted
+    ]},
+    notify={"webhook": "https://your-endpoint"},   # fires if IssueLoop itself breaks
+)
+
+issueloop.scan_repo("repos/myrepo")
+issueloop.run_tests("myrepo")
+issueloop.create_tickets("myrepo")
+
+ticket = issueloop.get_top_error("myrepo")   # claims one, marks in_progress
+issueloop.resolve(ticket["id"])              # your side fixed it
+```
+
+## Live monitoring
+
+Watch a process IssueLoop spawns itself, or tail a log file an
+already-running process writes to:
+
+```python
+from issueloop import watch_process, watch_log_file, stop_watch
+
+handle = watch_process("myrepo", "python3 main.py", cwd="/path/to/repo", debounce_seconds=3.0)
+# ...
+stop_watch(handle)
+```
+
+Or from the CLI, in the foreground:
+
+```bash
+issueloop watch myrepo --command "python3 main.py" --cwd /path/to/repo --debounce 3.0
+issueloop watch myrepo --log-file /path/to/app.log --debounce 3.0
+```
+
+Errors are debounced (default 3s of quiet) so a single multi-line
+traceback becomes one ticket, not dozens. Detected errors land in the
+same log format the batch test runner uses — `issueloop tickets` picks
+them up with no extra steps.
+
+## Fix-apply layer
+
+Off by default — every command needs an explicit allowlist entry:
+
+```yaml
+# config/permission.yaml
+per_repo:
+  myrepo:
+    allowed_patterns:
+      - 'sed -i .* somefile\.py'
+```
+
+```python
+issueloop.propose_fix(ticket_id, "sed -i 's/old/new/' somefile.py")
+result = issueloop.apply_fix(ticket_id)
+# {"status": "resolved" | "retry" | "escalated" | "denied", ...}
+```
+
+`apply_fix` re-runs the ticket's associated test after applying the
+fix and only resolves it if the test actually passes. Failures retry
+up to a configurable limit, then escalate to `needs_human`. Every
+permission decision is audited — `issueloop.get_permission_audit_log()`.
+
+## Use it from the CLI
+
+```bash
+issueloop check-env
+issueloop scan repos/myrepo
+issueloop test myrepo
+issueloop tickets myrepo
+issueloop next myrepo
+issueloop watch myrepo --command "..." | --log-file ...
+issueloop cleanup --days 30
+```
+
+## Use it from Node, Go, or anything else
+
+```bash
+issueloop serve --port 8787
+```
+
+```js
+const r = await fetch("http://localhost:8787/errors/top?repo=myrepo");
+const ticket = await r.json();
+```
+
+`serve` is localhost-only, no auth — put it behind your own gateway if
+you ever expose it further.
+
+## Full API
+
+Bug query — `get_all_bugs`, `get_unresolved_bugs`, `get_resolved_bugs`,
+`get_failed_bugs`, `get_bugs_needing_human`, `get_bugs_by_status`,
+`get_bugs_by_priority`, `get_bug`, `get_bug_count`,
+`get_bug_count_by_status`, `search_bugs`, `get_oldest_bug`, `get_newest_bug`
+
+Lifecycle — `get_top_error`, `get_all_errors`, `resolve`, `fail`,
+`escalate`, `reassign`, `retry_bug`, `get_bug_attempts`, `bulk_resolve`
+
+Fix-apply — `propose_fix`, `apply_fix`, `check_permission`,
+`get_permission_audit_log`
+
+Scan / test — `scan_repo`, `get_file_inventory`, `run_tests`,
+`run_single_test`, `create_tickets`
+
+Live monitoring — `watch_process`, `watch_log_file`, `stop_watch`,
+`list_active_watchers`
+
+LLM / tokens — `get_token_consumption`,
+`get_token_consumption_by_provider`, `get_llm_call_history`,
+`get_llm_provider_status`
+
+Database — `cleanup`, `purge_repo`, `get_database_stats`,
+`export_bugs`, `reap_stale_bugs`, `rotate_logs`
+
+Notifications — `get_crash_log`, `get_notification_config`
+
+Config / utility — `use`, `get_config`, `list_repos`, `health_check`
+
+## Config — every option
+
+| Field                 | Values                                | Default              |
+| ---------------------- | -------------------------------------- | --------------------- |
+| `database`             | `"local"`, `"supabase"`                | `"local"`             |
+| `database_path`        | any path                               | `data/issueloop.db`   |
+| `retention_days`       | int                                    | `30`                  |
+| `llm.providers`        | list of provider dicts, priority order | single ollama default |
+| `llm.provider`         | `"ollama"`, `"anthropic"`, `"openai"`  | `"ollama"`             |
+| `llm.apiKey`           | string                                 | none (required for anthropic/openai) |
+| `llm.tokenSize`        | int                                    | `1024`                |
+| `notify.webhook`       | URL                                    | none                  |
+
+Both `apiKey`/`api_key` and `tokenSize`/`token_size` work — camelCase
+or snake_case, your call. A single `llm={...}` dict still works exactly
+as before; wrap multiple in `llm={"providers": [...]}` for automatic
+fallback.
+
+Config files (`config/permission.yaml`, `config/provider_config.yaml`)
+resolve in this order: an explicit env var
+(`ISSUELOOP_PERMISSION_PATH` / `ISSUELOOP_PROVIDER_CONFIG_PATH`) →
+`./config/` relative to your current directory → `config/` in the
+IssueLoop checkout → bundled package defaults. This means a real
+`pip install` (not just `-e .`) still works correctly even without a
+project-local `config/` directory.
+
+## Why trust this
+
+- Every module has a test, and the tests run against real behavior —
+  a real SQLite file, a real HTTP server on a real socket — not just
+  mocks pretending things work.
+- Real bugs were caught and fixed by actually running this code, not
+  by reading it — including a prompt that only ever sent `stderr`
+  to the LLM triage step while pytest (and most test runners) report
+  failures on `stdout`, and a config-bundling gap that made the
+  permission system silently deny everything under a real (non-editable)
+  install. Both have regression coverage now.
+- Local-first by default. Nothing phones home, nothing requires an
+  account to try.
+- Read the code — it's small enough to actually read. That's on
+  purpose. File-by-file build order: `BUILD_ORDER.md`.
+
+## Testing
+
+```bash
+pytest tests/ -v
+```
+
+Full walkthrough, including live monitoring and the fix-apply layer: `TESTING.md`.
+
+## License
+
+MIT — see `LICENSE`.

issueloop-0.2.0/issueloop/__init__.py

@@ -0,0 +1,71 @@
+
+from pathlib import Path
+from typing import Optional
+
+from .config import use, get_config
+from .db import get_backend
+from .agent_state import TaskStatus
+from . import folder_reader, test_runner,ticket_creator
+
+
+def _backend():
+    cfg = get_config()
+    return get_backend(cfg.database, path=cfg.database_path)
+
+
+def scan_repo(repo_path: str):
+    out_file = folder_reader.write_inventory(Path(repo_path).resolve())
+    import json
+    return json.loads(out_file.read_text())
+
+
+def run_tests(repo_name: str):
+    return test_runner.run_tests(repo_name)
+
+
+def create_tickets(repo_name: str):
+    tickets = ticket_creator.create_tickets_for_repo(repo_name)
+    return [_ticket_to_dict(t) for t in tickets]
+
+
+def get_top_error(repo_name: str):
+    ticket = _backend().dispense_next(repo_name)
+    return _ticket_to_dict(ticket) if ticket else None
+
+
+def get_all_errors(repo_name: Optional[str] = None):
+    return [_ticket_to_dict(t) for t in _backend().get_open_tickets(repo_name)]
+
+
+def resolve(ticket_id: str):
+    _backend().update_ticket(ticket_id, status=TaskStatus.DONE)
+
+
+def fail(ticket_id: str):
+    _backend().update_ticket(ticket_id, status=TaskStatus.FAILED)
+
+
+def cleanup(older_than_days: Optional[int] = None, repo: Optional[str] = None):
+    days = older_than_days if older_than_days is not None else get_config().retention_days
+    return _backend().purge_old(days, repo)
+
+
+def _ticket_to_dict(ticket):
+    return {
+        "id": ticket.id,
+        "repo": ticket.repo,
+        "priority": ticket.priority.value,
+        "status": ticket.status.value,
+        "error_summary": ticket.error_summary,
+        "raw_log_ref": ticket.raw_log_ref,
+        "command": ticket.command,
+        "test_id": ticket.test_id,
+        "created_at": str(ticket.created_at),
+    }
+
+
+__all__ = [
+    "use", "get_config",
+    "scan_repo", "run_tests", "create_tickets",
+    "get_top_error", "get_all_errors", "resolve", "fail", "cleanup",
+]

issueloop-0.2.0/issueloop/agent_state.py

@@ -0,0 +1,75 @@
+from dataclasses import dataclass, field
+from datetime import datetime, timezone
+from enum import Enum
+from typing import Any, Optional
+
+
+class TaskStatus(str, Enum):
+    PENDING = "pending"
+    IN_PROGRESS = "in_progress"
+    BLOCKED = "blocked"
+    DONE = "done"
+    FAILED = "failed"
+    NEEDS_HUMAN = "needs_human"
+
+
+class TicketPriority(str, Enum):
+    BLOCKING = "blocking"
+    HIGH = "high"
+    NORMAL = "normal"
+    LOW = "low"
+
+
+@dataclass
+class SubTask:
+    id: str
+    description: str
+    depends_on: list[str] = field(default_factory=list)
+    status: TaskStatus = TaskStatus.PENDING
+    result: Optional[dict[str, Any]] = None
+
+
+@dataclass
+class Ticket:
+    id: str
+    repo: str
+    error_summary: str
+    raw_log_ref: str
+    priority: TicketPriority
+    status: TaskStatus = TaskStatus.PENDING
+    attempts: int = 0
+    created_at: datetime = field(default_factory=lambda: datetime.now(timezone.utc))
+    resolved_at: Optional[datetime] = None
+    escalation_summary: Optional[str] = None
+    command: Optional[str] = None
+    test_id: Optional[str] = None
+
+
+@dataclass
+class AgentState:
+    allowed_commands: list[str] = field(default_factory=list)
+    create_pr_permission: bool = False
+    supabase_write_permission: bool = False
+    repo: str = ""
+    branch: str = ""
+    current_task_id: str = ""
+    current_task: Optional[SubTask] = None
+    previous_task_and_result: Optional[dict[str, Any]] = None
+    next_task_id: Optional[str] = None
+    queue_independent_tasks: list[SubTask] = field(default_factory=list)
+    queue_dependent_task: list[SubTask] = field(default_factory=list)
+    active_tickets: list[Ticket] = field(default_factory=list)
+    message_from_user: Optional[str] = None
+    pending_confirmation: bool = False
+    goal: str = ""
+    what_to_do: list[str] = field(default_factory=list)
+    what_not_to_do: list[str] = field(default_factory=list)
+    overall_points: int = 0
+    current_task_points: int = 0
+    max_retries_per_ticket: int = 3
+    log_cache_ref: str = ""
+    last_error: Optional[str] = None
+
+
+def new_run_state(repo: str, goal: str, allowed_commands: list[str]):
+    return AgentState(repo=repo, goal=goal, allowed_commands=allowed_commands)