feat: add pi runner (#24)

This commit is contained in:
banteg
2026-01-02 16:13:55 +04:00
committed by GitHub
parent 7e5d6e3d40
commit d9c53b9e3a
11 changed files with 1098 additions and 8 deletions
+14 -4
View File
@@ -49,7 +49,7 @@ The orchestrator module containing:
- `/cancel` routes by reply-to progress message id (accepts extra text)
- `/{engine}` on the first line selects the engine for new threads
- Progress edits are throttled to 2s intervals and only run when new events arrive
- Resume tokens are runner-formatted command lines (e.g., `` `codex resume <token>` ``)
- Resume tokens are runner-formatted command lines (e.g., `` `codex resume <token>` ``, `` `claude --resume <token>` ``, `` `pi --session <path>` ``)
- Resume parsing polls all runners via `AutoRouter.resolve_resume()` and routes to the first match
- Bot command menu is synced on startup (`cancel` + engine commands)
@@ -92,6 +92,13 @@ The orchestrator module containing:
- Stderr is drained into a bounded tail (debug logging only)
- Translation errors abort the run; keep event normalization defensive
### `runners/pi.py` - Pi runner
| Component | Purpose |
|-----------|---------|
| `PiRunner` | Spawns `pi --print --mode json`, streams JSONL, emits takopi events |
| `translate_pi_event()` | Normalizes Pi JSONL into the takopi event schema |
### `model.py` / `runner.py` - Core domain types
| File | Purpose |
@@ -113,6 +120,8 @@ Auto-discovers runner modules in `takopi.runners` that export `BACKEND`.
| File | Purpose |
|------|---------|
| `codex.py` | Codex runner (JSONL → takopi events) + per-resume locks |
| `claude.py` | Claude runner (JSONL → takopi events) + per-resume locks |
| `pi.py` | Pi runner (JSONL → takopi events) + per-resume locks |
| `mock.py` | Mock runner for tests/demos |
### `config.py` - Configuration loading
@@ -166,7 +175,7 @@ handle_message() spawned as task with selected runner
Send initial progress message (silent)
runner.run(prompt, resume_token)
├── Spawns engine subprocess (e.g., codex exec --json)
├── Spawns engine subprocess (e.g., codex exec --json, pi --print --mode json)
├── Streams JSONL from stdout
├── Normalizes JSONL -> takopi events
├── Yields Takopi events (async iterator)
@@ -184,8 +193,8 @@ Send/edit final message
### Resume Flow
Same as above; auto-router polls all runners to extract resume tokens:
- Router returns first matching token (e.g. `` `claude --resume <id>` `` routes to Claude)
- Selected runner spawns with resume (e.g. `codex exec --json resume <token> -`)
- Router returns first matching token (e.g. `` `claude --resume <id>` `` routes to Claude, `` `pi --session <path>` `` routes to Pi)
- Selected runner spawns with resume (e.g. `codex exec --json resume <token> -`, `pi --print --mode json --session <path> <prompt>`)
- Per-token lock serializes concurrent resumes on the same thread
## Error Handling
@@ -193,6 +202,7 @@ Same as above; auto-router polls all runners to extract resume tokens:
| Scenario | Behavior |
|----------|----------|
| `codex exec` fails (rc != 0) | Emits a warning `action` plus `completed(ok=false, error=...)` |
| `pi` fails (rc != 0) | Emits a warning `action` plus `completed(ok=false, error=...)` |
| Telegram API error | Logged, edit skipped (progress continues) |
| Cancellation | Cancel scope terminates the process group (POSIX) and renders `cancelled` |
| Errors in handler | Final render uses `status=error` and preserves resume tokens when known |
+137
View File
@@ -0,0 +1,137 @@
Below is a concrete implementation spec for adding **Pi (pi-coding-agent CLI)** as a first-class engine in Takopi (v0.4.0).
---
## Scope
### Goal
Add a new engine backend **`pi`** so Takopi can:
* Run Pi non-interactively via the **pi CLI** (`pi --print`).
* Stream progress by parsing **`--mode json`** (newline-delimited JSON). Each line is a JSON object.
* Support resumable sessions via **`--session <path>`** (Takopi emits a canonical resume line the user can reply with).
### Non-goals (v1)
* Interactive TUI flows (session picker, prompts, etc.)
* RPC mode (requires a long-running process and JSON commands)
---
## UX and behavior
### Engine selection
* Existing: `takopi codex`
* New: `takopi pi`
### Resume UX (canonical line)
Takopi appends a **single backticked** resume line at the end of the message, like:
```text
`pi --session /home/user/.pi/agent/sessions/--repo--/2026-01-02T12-34-56-789Z_abcd.jsonl`
```
Notes:
* `pi --resume/-r` opens an interactive session picker, so Takopi uses `--session <path>` instead.
* The resume token is the **session file path** (JSONL), treated as an opaque string.
* If the path contains spaces, the runner will quote it.
### Non-interactive runs
Use `--print` and `--mode json` for headless JSONL output.
Pi does not accept `-- <prompt>` to protect prompts starting with `-`. Takopi prefixes a leading space if the prompt begins with `-` so it is not parsed as a flag.
---
## Config additions
Takopi config lives at either:
* `.takopi/takopi.toml` (project-local), or
* `~/.takopi/takopi.toml` (home).
Add a new optional `[pi]` section.
Recommended v1 schema:
```toml
# .takopi/takopi.toml
default_engine = "pi"
[pi]
cmd = "pi" # optional; defaults to "pi"
extra_args = [] # optional list of strings, appended verbatim
model = "..." # optional; passed as --model
provider = "..." # optional; passed as --provider
session_dir = "..." # optional; directory for session files
session_title = "pi" # optional; defaults to model or "pi"
```
Notes:
* `extra_args` lets you pass new Pi flags without changing Takopi.
* If `session_dir` is omitted, Takopi uses Pi's default session dir:
`~/.pi/agent/sessions/--<cwd>--` (with path separators replaced by `-`).
---
## Code changes (by file)
### 1) New file: `src/takopi/runners/pi.py`
Expose a module-level `BACKEND = EngineBackend(...)`.
#### Runner invocation
The runner should launch Pi in headless JSON mode:
```text
pi --print --mode json --session <session.jsonl> <prompt>
```
When resuming, `<session.jsonl>` is the resume token extracted from the chat.
#### Event translation
Pi JSONL output is `AgentSessionEvent` (from `@mariozechner/pi-agent-core`).
The runner should translate:
* `tool_execution_start` -> `action` (phase: started)
* `tool_execution_end` -> `action` (phase: completed)
* `agent_end` -> `completed`
For the final answer, use the most recent assistant message text (from
`message_end` events). For errors, if the assistant stopReason is `error` or
`aborted`, emit `completed(ok=false, error=...)`.
---
## Installation and auth
Install the CLI globally:
```text
npm install -g @mariozechner/pi-coding-agent
```
Auth is stored under `~/.pi/agent/auth.json`. Run `pi` once interactively to
set up credentials before using Takopi.
---
## Known pitfalls
* `--resume` is interactive; Takopi uses `--session <path>` instead.
* Prompts that start with `-` are interpreted as flags by the CLI. Takopi
prefixes a space to make them safe.
---
If you want, I can also add a sample `takopi.toml` snippet to the README or
include a small quickstart section for Pi in the onboarding panel.
@@ -0,0 +1,67 @@
# Pi `--mode json` event cheatsheet
`pi --print --mode json` writes **one JSON object per line** (JSONL) with a
required `type` field. These are `AgentSessionEvent` objects from
`@mariozechner/pi-agent-core`.
## Top-level event lines
### `agent_start`
```json
{"type":"agent_start"}
```
### `agent_end`
```json
{"type":"agent_end","messages":[{"role":"assistant","content":[{"type":"text","text":"Done."}],"stopReason":"stop","timestamp":123}]}
```
### `turn_start` / `turn_end`
```json
{"type":"turn_start"}
```
```json
{"type":"turn_end","message":{...},"toolResults":[...]}
```
### `message_start` / `message_update` / `message_end`
```json
{"type":"message_start","message":{"role":"assistant","content":[{"type":"text","text":"Working..."}]}}
```
```json
{"type":"message_update","message":{...},"assistantMessageEvent":{"type":"text_delta","delta":"...","contentIndex":0}}
```
```json
{"type":"message_end","message":{"role":"assistant","content":[{"type":"text","text":"Done."}],"stopReason":"stop"}}
```
### `tool_execution_start`
```json
{"type":"tool_execution_start","toolCallId":"tool_1","toolName":"bash","args":{"command":"ls"}}
```
### `tool_execution_update`
```json
{"type":"tool_execution_update","toolCallId":"tool_1","toolName":"bash","args":{"command":"ls"},"partialResult":{"content":[{"type":"text","text":"..."}]}}
```
### `tool_execution_end`
```json
{"type":"tool_execution_end","toolCallId":"tool_1","toolName":"bash","result":{"content":[{"type":"text","text":"ok"}],"details":{}},"isError":false}
```
## Notes
* `message_end` with `role = "assistant"` contains the final assistant text.
* `assistantMessageEvent` in `message_update` provides streaming deltas.
* `tool_execution_*` events map cleanly to Takopi `action` events.
+154
View File
@@ -0,0 +1,154 @@
# Pi -> Takopi event mapping (spec)
This document specifies how to add a Pi runner to Takopi by translating
Pi CLI `--mode json` JSONL events into Takopi events. The Pi JSONL stream is
`AgentSessionEvent` from `@mariozechner/pi-agent-core`.
The goal is to make Pi feel identical to the Codex/Claude runners from the
bridge/renderer point of view while preserving Takopi invariants (stable action
ids, per-session serialization, single completed event).
---
## 1. Input stream contract (Pi CLI)
Pi CLI emits **one JSON object per line** (JSONL) when invoked with:
```
pi --print --mode json <prompt>
```
Notes:
- `--print` is required for non-interactive runs.
- `--mode json` outputs all agent events (no TUI banners).
- Pi does not support `-- <prompt>`; prompts starting with `-` must be
prefixed (Takopi does this automatically).
---
## 2. Resume tokens and resume lines
- Engine id: `pi`
- Canonical resume line (embedded in chat):
```
`pi --session <path>`
```
The token is the **session JSONL file path**.
Why not `--resume`?
- `--resume/-r` opens an interactive session picker; it does not accept a
session token. Takopi must use `--session <path>` instead.
---
## 3. Session lifecycle + serialization
Takopi requires **serialization per session token**:
- For new runs (`resume=None`), do **not** acquire a lock until a `started`
event is emitted (Takopi emits this as soon as the first JSON event arrives).
- Once the session is known, acquire a lock for `pi:<session_path>` and hold it
until the run completes.
- For resumed runs, acquire the lock immediately on entry.
---
## 4. Event translation (Pi JSONL -> Takopi)
Pi emits `AgentSessionEvent` objects. Only a subset is required for Takopi.
### 4.1 `tool_execution_start`
Example:
```json
{"type":"tool_execution_start","toolCallId":"tool_1","toolName":"bash","args":{"command":"ls"}}
```
Mapping:
- Emit `action` with `phase="started"`.
- `action.id = toolCallId`.
- `action.kind` from tool name (see section 5).
- `action.title` derived from tool + args.
### 4.2 `tool_execution_end`
Example:
```json
{"type":"tool_execution_end","toolCallId":"tool_1","toolName":"bash","result":{...},"isError":false}
```
Mapping:
- Emit `action` with `phase="completed"`.
- `ok = !isError`.
- Carry `result` and `isError` in `detail` for debugging.
### 4.3 `message_end` (assistant)
Pi emits message lifecycle events. For `message_end` where `message.role == "assistant"`:
- Store the latest assistant text as the **final answer fallback**.
- If `stopReason` is `error` or `aborted`, store `errorMessage`.
- Capture `usage` for `completed.usage`.
### 4.4 `agent_end`
Example:
```json
{"type":"agent_end","messages":[...]}
```
Mapping:
- Emit a single `completed` event:
- `ok = true` unless the last assistant message has `stopReason` `error` or `aborted`.
- `answer = last assistant text` (from `message_end` or `agent_end.messages`).
- `error = errorMessage` if present.
- `resume = ResumeToken(engine="pi", value=session_path)`.
- `usage = last assistant usage`.
### 4.5 Other events
Ignore unknown events. If a JSONL line is malformed, emit a warning action and
continue (default `JsonlSubprocessRunner` behavior).
---
## 5. Tool name -> ActionKind mapping heuristics
Pi tool names are lower-case by default. Suggested mapping:
| Tool name | ActionKind | Title logic |
| --- | --- | --- |
| `bash` | `command` | `args.command` |
| `edit`, `write` | `file_change` | `args.path` |
| `read` | `tool` | `read: <path>` |
| `grep` | `tool` | `grep: <pattern>` |
| `find` | `tool` | `find: <pattern>` |
| `ls` | `tool` | `ls: <path>` |
| (default) | `tool` | tool name |
For `file_change`, include `detail.changes = [{"path": <path>, "kind": "update"}]`.
---
## 6. Usage mapping
Takopi `completed.usage` should mirror Pi's assistant `usage` object without
transformation.
---
## 7. Suggested Takopi config keys
A minimal TOML config for Pi:
```toml
[pi]
cmd = "pi"
model = "..."
provider = "..."
extra_args = []
```
Use `extra_args` for any newer Pi CLI flags not explicitly mapped.
+3 -2
View File
@@ -23,7 +23,7 @@ Out of scope for v0.4.0:
## 2. Terminology
- **EngineId**: string identifier of an engine (e.g., `"codex"`).
- **EngineId**: string identifier of an engine (e.g., `"codex"`, `"claude"`, `"pi"`).
- **Runner**: Takopi adapter that executes an engine process and yields **Takopi events**.
- **Thread**: a single engine-side conversation, identified in Takopi by a **ResumeToken**.
- **ResumeToken**: Takopi-owned thread identifier `{ engine: EngineId, value: str }`.
@@ -41,6 +41,7 @@ The canonical ResumeLine embedded in chat MUST be the engines CLI resume comm
- `codex resume <id>`
- `claude --resume <id>`
- `pi --session <path>`
Takopi MUST treat the runner as authoritative for:
@@ -347,7 +348,7 @@ Decision (v0.4.0):
* If an engine subcommand is provided, Takopi MUST still use the auto-router, but it overrides the configured default engine for new threads.
* Resume extraction MUST poll **all** available runners (per §3.4) and route to the first matching runner.
* New thread engine override (chat-level):
* Users MAY prefix the first non-empty line with `/{engine}` (e.g. `/claude` or `/codex`) to select the engine for a **new** thread.
* Users MAY prefix the first non-empty line with `/{engine}` (e.g. `/claude`, `/codex`, or `/pi`) to select the engine for a **new** thread.
* The bridge MUST strip that directive from the prompt before invoking the runner.
* If a ResumeToken is resolved from the message or reply, it MUST take precedence and the `/{engine}` directive MUST be ignored.