pocketcoder-a1 0.1.0__tar.gz

pocketcoder_a1-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Dmitry Chashchin
+
+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.

pocketcoder_a1-0.1.0/PKG-INFO

@@ -0,0 +1,230 @@
+Metadata-Version: 2.4
+Name: pocketcoder-a1
+Version: 0.1.0
+Summary: Autonomous Coding Agent — The Autonomous Gnome
+Author-email: Dmitry Chashchin <chashchin.dmitry@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/Chashchin-Dmitry/pocketcoder-a1
+Project-URL: Repository, https://github.com/Chashchin-Dmitry/pocketcoder-a1
+Keywords: ai,coding,agent,autonomous,claude,llm
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development
+Classifier: Topic :: Software Development :: Code Generators
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: ruff>=0.1; extra == "dev"
+Provides-Extra: api
+Requires-Dist: anthropic>=0.18; extra == "api"
+Provides-Extra: ollama
+Requires-Dist: ollama>=0.1; extra == "ollama"
+Provides-Extra: test
+Requires-Dist: playwright>=1.40; extra == "test"
+Dynamic: license-file
+
+# PocketCoder-A1
+
+**Autonomous Coding Agent with Web Dashboard**
+
+A1 works on your tasks autonomously — you add tasks, start the agent, watch progress in the dashboard.
+
+---
+
+## Quick Start (5 steps)
+
+```bash
+# 1. Clone and install
+git clone https://github.com/Chashchin-Dmitry/pocketcoder-a1.git
+cd pocketcoder-a1
+python3 -m venv .venv
+source .venv/bin/activate
+pip install -e .
+
+# 2. Install browser for E2E tests (optional)
+pip install playwright requests
+playwright install chromium
+
+# 3. Initialize in any project
+pca init /path/to/your-project
+
+# 4. Add tasks
+pca task add "Add login page" -d /path/to/your-project
+pca task add "Write unit tests" -d /path/to/your-project
+
+# 5. Launch dashboard
+pca ui -d /path/to/your-project
+# Opens http://localhost:7331
+```
+
+**Requirements:** Python 3.10+, [Claude Code CLI](https://docs.anthropic.com/en/docs/claude-code) installed (`npm i -g @anthropic-ai/claude-code`)
+
+---
+
+## Web Dashboard
+
+Open `http://localhost:7331` after running `pca ui`. 7 pages:
+
+| Page | What it does |
+|------|-------------|
+| **Dashboard** | Status cards, task list, Quick Add form, Start/Stop agent, Live Log |
+| **Tasks** | All tasks with priority badges (#1 #2 #3), drag-and-drop reorder, Add Task + Add Thought forms |
+| **Sessions** | Current and previous session details |
+| **Activity Log** | Timeline of all actions (started, stopped, task added) |
+| **Commits** | Git commit history |
+| **Transform** | Paste raw text -> AI breaks it into structured tasks |
+| **Settings** | Theme toggle (dark/light), provider info |
+
+### Key features:
+- **Live Agent Log** — see what Claude is doing in real-time (action icons + raw log)
+- **Queue Message** — send instructions to running agent (appears when agent is Running)
+- **Drag-and-Drop** — reorder task priorities on Tasks page
+- **Dark Theme** — click moon icon top-right
+- **Transform** — paste messy notes, AI creates structured tasks with checkboxes
+
+---
+
+## How to Test (step by step)
+
+### Option A: Use the dashboard (recommended)
+
+```bash
+source .venv/bin/activate
+
+# Start dashboard on any project
+pca ui -d /path/to/your-project
+```
+
+Then in browser at `http://localhost:7331`:
+
+1. **Add tasks** — Quick Add form on Dashboard, or Tasks page
+2. **Start Agent** — green button, watch status change to "Running"
+3. **Watch Live Log** — "Agent Live Log" panel shows what Claude does
+4. **Send message** — "Message to Agent" form appears when running
+5. **Stop** — red "Stop Agent" button
+6. **Transform** — go to Transform page, paste text, click "AI Transform"
+7. **Check results** — Tasks page shows green checkmarks for completed tasks
+
+### Option B: CLI only
+
+```bash
+source .venv/bin/activate
+
+pca init ./my-project
+pca task add "Create hello.py with greet function" -d ./my-project
+pca start -d ./my-project          # Runs agent in terminal
+pca status -d ./my-project         # Check progress
+pca tasks -d ./my-project          # See task list
+```
+
+### Option C: Run E2E tests
+
+```bash
+source .venv/bin/activate
+
+# Start dashboard in background
+pca ui -d ./my-project --no-browser &
+
+# Run vision-based tests (7 scenarios)
+pca test -d ./my-project --no-vision
+
+# Reports saved to .a1/test-reports/
+```
+
+---
+
+## CLI Commands
+
+| Command | Description |
+|---------|-------------|
+| `pca init <dir>` | Initialize .a1/ in project |
+| `pca task add "..."` | Add task with optional description |
+| `pca think "..."` | Add raw thought/idea |
+| `pca tasks` | Show all tasks |
+| `pca start` | Start autonomous agent (CLI mode) |
+| `pca status` | Show current status |
+| `pca validate` | Run validation (syntax, tests, lint) |
+| `pca ui` | Launch web dashboard on :7331 |
+| `pca test` | Run E2E vision tests |
+| `pca log` | Show session history |
+
+Add `-d /path/to/project` to any command to specify project directory.
+
+---
+
+## Dashboard API (17 endpoints)
+
+| Method | Endpoint | What it does |
+|--------|----------|-------------|
+| GET | `/` | Dashboard page |
+| GET | `/tasks` | Tasks page |
+| GET | `/sessions` | Sessions page |
+| GET | `/log` | Activity log page |
+| GET | `/commits` | Git commits page |
+| GET | `/transform` | Transform page |
+| GET | `/settings` | Settings page |
+| GET | `/api/status` | JSON: checkpoint + tasks + progress + running |
+| GET | `/api/log?since=N` | JSON: agent log entries since index N |
+| POST | `/add-task` | Add task (form: task, description) |
+| POST | `/add-thought` | Add thought (form: thought) |
+| POST | `/start` | Start agent |
+| POST | `/stop` | Stop agent |
+| POST | `/queue-message` | Message to agent (form: message) |
+| POST | `/api/reorder` | Reorder tasks (JSON: {order: [ids]}) |
+| POST | `/transform` | AI transform text to tasks (form: text) |
+| POST | `/transform-confirm` | Confirm transformed tasks (JSON: {tasks}) |
+
+---
+
+## Project Structure
+
+```
+pocketcoder-a1/
+├── a1/                      # Main code
+│   ├── checkpoint.py        # State between sessions
+│   ├── tasks.py             # Task management (priorities, reorder)
+│   ├── validator.py         # Validation (syntax, tests, lint)
+│   ├── loop.py              # Agent session loop (Claude CLI subprocess)
+│   ├── dashboard.py         # Web UI (17 endpoints, inline HTML/CSS/JS)
+│   ├── cli.py               # CLI commands (pca)
+│   └── tester/              # Vision-based QA agent
+├── .a1/                     # Data (created on pca init)
+│   ├── checkpoint.json      # Current state
+│   ├── tasks.json           # Task list with priorities
+│   ├── queue.json           # Message queue for agent
+│   ├── sessions/            # Session logs
+│   └── checkpoints/         # Checkpoint archive
+├── sandbox/                 # Test projects
+│   ├── test-e2e/            # E2E test (3/3 tasks passed)
+│   └── epotos-templates/    # Real project test (3/3 tasks passed)
+├── CLAUDE.md                # Agent instructions
+├── CURRENT_STAGE.md         # Status with cause-effect chains
+└── pyproject.toml           # pip install -e .
+```
+
+---
+
+## Providers
+
+| Provider | Command | Requires |
+|----------|---------|----------|
+| claude-max | `pca start` | Claude Code CLI + Max subscription |
+| claude-api | `pca start --provider claude-api` | ANTHROPIC_API_KEY |
+| ollama | `pca start --provider ollama` | Local Ollama server |
+
+---
+
+## License
+
+MIT

pocketcoder_a1-0.1.0/README.md

@@ -0,0 +1,194 @@
+# PocketCoder-A1
+
+**Autonomous Coding Agent with Web Dashboard**
+
+A1 works on your tasks autonomously — you add tasks, start the agent, watch progress in the dashboard.
+
+---
+
+## Quick Start (5 steps)
+
+```bash
+# 1. Clone and install
+git clone https://github.com/Chashchin-Dmitry/pocketcoder-a1.git
+cd pocketcoder-a1
+python3 -m venv .venv
+source .venv/bin/activate
+pip install -e .
+
+# 2. Install browser for E2E tests (optional)
+pip install playwright requests
+playwright install chromium
+
+# 3. Initialize in any project
+pca init /path/to/your-project
+
+# 4. Add tasks
+pca task add "Add login page" -d /path/to/your-project
+pca task add "Write unit tests" -d /path/to/your-project
+
+# 5. Launch dashboard
+pca ui -d /path/to/your-project
+# Opens http://localhost:7331
+```
+
+**Requirements:** Python 3.10+, [Claude Code CLI](https://docs.anthropic.com/en/docs/claude-code) installed (`npm i -g @anthropic-ai/claude-code`)
+
+---
+
+## Web Dashboard
+
+Open `http://localhost:7331` after running `pca ui`. 7 pages:
+
+| Page | What it does |
+|------|-------------|
+| **Dashboard** | Status cards, task list, Quick Add form, Start/Stop agent, Live Log |
+| **Tasks** | All tasks with priority badges (#1 #2 #3), drag-and-drop reorder, Add Task + Add Thought forms |
+| **Sessions** | Current and previous session details |
+| **Activity Log** | Timeline of all actions (started, stopped, task added) |
+| **Commits** | Git commit history |
+| **Transform** | Paste raw text -> AI breaks it into structured tasks |
+| **Settings** | Theme toggle (dark/light), provider info |
+
+### Key features:
+- **Live Agent Log** — see what Claude is doing in real-time (action icons + raw log)
+- **Queue Message** — send instructions to running agent (appears when agent is Running)
+- **Drag-and-Drop** — reorder task priorities on Tasks page
+- **Dark Theme** — click moon icon top-right
+- **Transform** — paste messy notes, AI creates structured tasks with checkboxes
+
+---
+
+## How to Test (step by step)
+
+### Option A: Use the dashboard (recommended)
+
+```bash
+source .venv/bin/activate
+
+# Start dashboard on any project
+pca ui -d /path/to/your-project
+```
+
+Then in browser at `http://localhost:7331`:
+
+1. **Add tasks** — Quick Add form on Dashboard, or Tasks page
+2. **Start Agent** — green button, watch status change to "Running"
+3. **Watch Live Log** — "Agent Live Log" panel shows what Claude does
+4. **Send message** — "Message to Agent" form appears when running
+5. **Stop** — red "Stop Agent" button
+6. **Transform** — go to Transform page, paste text, click "AI Transform"
+7. **Check results** — Tasks page shows green checkmarks for completed tasks
+
+### Option B: CLI only
+
+```bash
+source .venv/bin/activate
+
+pca init ./my-project
+pca task add "Create hello.py with greet function" -d ./my-project
+pca start -d ./my-project          # Runs agent in terminal
+pca status -d ./my-project         # Check progress
+pca tasks -d ./my-project          # See task list
+```
+
+### Option C: Run E2E tests
+
+```bash
+source .venv/bin/activate
+
+# Start dashboard in background
+pca ui -d ./my-project --no-browser &
+
+# Run vision-based tests (7 scenarios)
+pca test -d ./my-project --no-vision
+
+# Reports saved to .a1/test-reports/
+```
+
+---
+
+## CLI Commands
+
+| Command | Description |
+|---------|-------------|
+| `pca init <dir>` | Initialize .a1/ in project |
+| `pca task add "..."` | Add task with optional description |
+| `pca think "..."` | Add raw thought/idea |
+| `pca tasks` | Show all tasks |
+| `pca start` | Start autonomous agent (CLI mode) |
+| `pca status` | Show current status |
+| `pca validate` | Run validation (syntax, tests, lint) |
+| `pca ui` | Launch web dashboard on :7331 |
+| `pca test` | Run E2E vision tests |
+| `pca log` | Show session history |
+
+Add `-d /path/to/project` to any command to specify project directory.
+
+---
+
+## Dashboard API (17 endpoints)
+
+| Method | Endpoint | What it does |
+|--------|----------|-------------|
+| GET | `/` | Dashboard page |
+| GET | `/tasks` | Tasks page |
+| GET | `/sessions` | Sessions page |
+| GET | `/log` | Activity log page |
+| GET | `/commits` | Git commits page |
+| GET | `/transform` | Transform page |
+| GET | `/settings` | Settings page |
+| GET | `/api/status` | JSON: checkpoint + tasks + progress + running |
+| GET | `/api/log?since=N` | JSON: agent log entries since index N |
+| POST | `/add-task` | Add task (form: task, description) |
+| POST | `/add-thought` | Add thought (form: thought) |
+| POST | `/start` | Start agent |
+| POST | `/stop` | Stop agent |
+| POST | `/queue-message` | Message to agent (form: message) |
+| POST | `/api/reorder` | Reorder tasks (JSON: {order: [ids]}) |
+| POST | `/transform` | AI transform text to tasks (form: text) |
+| POST | `/transform-confirm` | Confirm transformed tasks (JSON: {tasks}) |
+
+---
+
+## Project Structure
+
+```
+pocketcoder-a1/
+├── a1/                      # Main code
+│   ├── checkpoint.py        # State between sessions
+│   ├── tasks.py             # Task management (priorities, reorder)
+│   ├── validator.py         # Validation (syntax, tests, lint)
+│   ├── loop.py              # Agent session loop (Claude CLI subprocess)
+│   ├── dashboard.py         # Web UI (17 endpoints, inline HTML/CSS/JS)
+│   ├── cli.py               # CLI commands (pca)
+│   └── tester/              # Vision-based QA agent
+├── .a1/                     # Data (created on pca init)
+│   ├── checkpoint.json      # Current state
+│   ├── tasks.json           # Task list with priorities
+│   ├── queue.json           # Message queue for agent
+│   ├── sessions/            # Session logs
+│   └── checkpoints/         # Checkpoint archive
+├── sandbox/                 # Test projects
+│   ├── test-e2e/            # E2E test (3/3 tasks passed)
+│   └── epotos-templates/    # Real project test (3/3 tasks passed)
+├── CLAUDE.md                # Agent instructions
+├── CURRENT_STAGE.md         # Status with cause-effect chains
+└── pyproject.toml           # pip install -e .
+```
+
+---
+
+## Providers
+
+| Provider | Command | Requires |
+|----------|---------|----------|
+| claude-max | `pca start` | Claude Code CLI + Max subscription |
+| claude-api | `pca start --provider claude-api` | ANTHROPIC_API_KEY |
+| ollama | `pca start --provider ollama` | Local Ollama server |
+
+---
+
+## License
+
+MIT

pocketcoder_a1-0.1.0/a1/__init__.py

@@ -0,0 +1,6 @@
+"""
+PocketCoder-A1: Autonomous Coding Agent
+"""
+
+__version__ = "0.1.0"
+__codename__ = "Autonomous Gnome"

pocketcoder_a1-0.1.0/a1/checkpoint.py

@@ -0,0 +1,146 @@
+"""
+Checkpoint Manager — сохранение состояния между сессиями
+"""
+
+import json
+from datetime import datetime
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+
+
+class CheckpointManager:
+    """Управление checkpoint'ами для автономной работы"""
+
+    def __init__(self, project_dir: Path):
+        self.project_dir = Path(project_dir)
+        self.a1_dir = self.project_dir / ".a1"
+        self.checkpoint_file = self.a1_dir / "checkpoint.json"
+        self.checkpoints_dir = self.a1_dir / "checkpoints"
+
+        # Создаём структуру если не существует
+        self.a1_dir.mkdir(parents=True, exist_ok=True)
+        self.checkpoints_dir.mkdir(exist_ok=True)
+
+    def load(self) -> Dict[str, Any]:
+        """Загрузить текущий checkpoint"""
+        if not self.checkpoint_file.exists():
+            return self._create_initial()
+
+        try:
+            with open(self.checkpoint_file, "r") as f:
+                return json.load(f)
+        except (json.JSONDecodeError, IOError):
+            return self._create_initial()
+
+    def save(self, checkpoint: Dict[str, Any]) -> None:
+        """Сохранить checkpoint"""
+        checkpoint["updated_at"] = datetime.now().isoformat()
+
+        # Сохраняем текущий
+        with open(self.checkpoint_file, "w") as f:
+            json.dump(checkpoint, f, indent=2, ensure_ascii=False)
+
+        # Архивируем копию
+        session = checkpoint.get("session", 0)
+        archive_file = self.checkpoints_dir / f"session_{session:03d}.json"
+        with open(archive_file, "w") as f:
+            json.dump(checkpoint, f, indent=2, ensure_ascii=False)
+
+    def _create_initial(self) -> Dict[str, Any]:
+        """Создать начальный checkpoint"""
+        return {
+            "status": "STARTING",
+            "session": 0,
+            "current_task": None,
+            "context_percent": 0,
+            "files_modified": [],
+            "decisions": [],
+            "next_steps": [],
+            "last_action": None,
+            "created_at": datetime.now().isoformat(),
+            "updated_at": datetime.now().isoformat(),
+        }
+
+    def start_session(self) -> Dict[str, Any]:
+        """Начать новую сессию"""
+        checkpoint = self.load()
+        checkpoint["session"] += 1
+        checkpoint["status"] = "WORKING"
+        checkpoint["session_started_at"] = datetime.now().isoformat()
+        self.save(checkpoint)
+        return checkpoint
+
+    def end_session(
+        self,
+        current_task: Optional[str] = None,
+        files_modified: Optional[List[str]] = None,
+        decisions: Optional[List[str]] = None,
+        next_steps: Optional[List[str]] = None,
+        last_action: Optional[str] = None,
+        context_percent: int = 0,
+    ) -> None:
+        """Завершить сессию с сохранением состояния"""
+        checkpoint = self.load()
+
+        if current_task:
+            checkpoint["current_task"] = current_task
+        if files_modified:
+            # Добавляем к существующим, убираем дубли
+            existing = set(checkpoint.get("files_modified", []))
+            existing.update(files_modified)
+            checkpoint["files_modified"] = list(existing)
+        if decisions:
+            checkpoint["decisions"].extend(decisions)
+            checkpoint["decisions"] = checkpoint["decisions"][-20:]
+        if next_steps:
+            checkpoint["next_steps"] = next_steps
+        if last_action:
+            checkpoint["last_action"] = last_action
+
+        checkpoint["context_percent"] = context_percent
+        checkpoint["session_ended_at"] = datetime.now().isoformat()
+
+        self.save(checkpoint)
+
+    def mark_completed(self) -> None:
+        """Отметить всю работу как завершённую"""
+        checkpoint = self.load()
+        checkpoint["status"] = "COMPLETED"
+        checkpoint["completed_at"] = datetime.now().isoformat()
+        self.save(checkpoint)
+
+    def is_completed(self) -> bool:
+        """Проверить завершена ли работа"""
+        checkpoint = self.load()
+        return checkpoint.get("status") == "COMPLETED"
+
+    def get_session_number(self) -> int:
+        """Получить номер текущей сессии"""
+        checkpoint = self.load()
+        return checkpoint.get("session", 0)
+
+    def get_summary(self) -> str:
+        """Получить текстовое резюме для промпта"""
+        cp = self.load()
+
+        lines = [
+            f"## Checkpoint (Session #{cp['session']})",
+            f"Status: {cp['status']}",
+            f"Current task: {cp.get('current_task', 'None')}",
+            f"Last action: {cp.get('last_action', 'None')}",
+        ]
+
+        if cp.get("files_modified"):
+            lines.append(f"Files modified: {', '.join(cp['files_modified'][-10:])}")
+
+        if cp.get("decisions"):
+            lines.append("Recent decisions:")
+            for d in cp["decisions"][-5:]:
+                lines.append(f"  - {d}")
+
+        if cp.get("next_steps"):
+            lines.append("Next steps:")
+            for s in cp["next_steps"]:
+                lines.append(f"  - {s}")
+
+        return "\n".join(lines)