priests 0.4.0__tar.gz

priests-0.4.0/.claude/settings.local.json

@@ -0,0 +1,18 @@
+{
+  "permissions": {
+    "allow": [
+      "Bash(wc:*)",
+      "Bash(uv run:*)",
+      "Bash(uv sync:*)",
+      "Bash(uv pip:*)",
+      "Bash(curl -s http://127.0.0.1:8777/v1/sessions)",
+      "Bash(python3 -m json.tool)",
+      "Bash(curl -s -X POST http://127.0.0.1:8777/v1/chat -H 'Content-Type: application/json' -d '{\"prompt\": \"Remember the number 42.\", \"session_id\": \"test-session-1\", \"no_think\": true}')",
+      "Bash(curl -s -X POST http://127.0.0.1:8777/v1/chat -H 'Content-Type: application/json' -d '{\"prompt\": \"What number did I ask you to remember?\", \"session_id\": \"test-session-1\", \"no_think\": true}')",
+      "Bash(curl -s http://127.0.0.1:8777/v1/sessions/test-session-1)",
+      "Bash(python -c ':*)",
+      "Bash(kill %1)",
+      "Bash(ls:*)"
+    ]
+  }
+}

priests-0.4.0/.gitignore

@@ -0,0 +1,208 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+#poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+#pdm.lock
+#pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+#pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Abstra
+# Abstra is an AI-powered process automation framework.
+# Ignore directories containing user credentials, local state, and settings.
+# Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#  and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#  you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/
+.DS_Store

priests-0.4.0/CHANGELOG.md

@@ -0,0 +1,84 @@
+# Changelog
+
+All notable changes to `priests` are documented here.
+
+## [0.4.0] — 2026-04-04
+
+### Added
+- Streaming output in CLI — `priests run` and interactive chat now print tokens
+  as they arrive using `engine.stream()`; time-to-first-token ~215ms vs ~3s wait
+  for buffered responses (14× perceived speedup on local models)
+
+---
+
+## [0.3.0] — 2026-04-04
+
+### Added
+- Provider registry (`priests/registry.py`) — 12 providers with curated model
+  lists: Ollama, OpenAI, Anthropic, Gemini, Bailian, Alibaba Cloud, MiniMax,
+  DeepSeek, Kimi, Groq, OpenRouter, Custom
+- `priests model list` — list saved provider/model pairs, marks current default
+- `priests model default` — interactive arrow-key selection of default model
+- `priests model add` — guided flow: select provider, enter API key, pick model
+- `priests providers` — table of all providers with configured/unconfigured status
+- `priests providers models <name>` — show curated model list or fetch Ollama
+  models dynamically
+- Proxy support — `[proxy] url` in `priests.toml`; `use_proxy = true/false` per
+  provider; engine resolves and passes proxy URL to adapter constructors
+- DeepSeek provider (`api.deepseek.com`) — `deepseek-chat`, `deepseek-reasoner`
+- Kimi / Moonshot provider (`api.moonshot.cn`) — K2.5, K2-thinking, moonshot-v1
+- `[models]` section in `priests.toml` — stores `provider/model` pairs
+
+### Fixed
+- Gemini base URL corrected to `/v1beta/openai/` (was `/openai/v1` → 404)
+- MiniMax base URL corrected to `api.minimax.io/v1` (was `api.minimax.chat/v1` → 401)
+- `think` parameter now only injected for Qwen/Ollama providers; other providers
+  (Gemini, OpenAI, etc.) receive empty `provider_options` to avoid HTTP 400
+
+### Changed
+- Provider labels updated: `(domestic)` → `(China mainland)`
+- Model lists refreshed: gpt-4.1 family, qwen3.5, kimi-k2.5, MiniMax M2.x,
+  Claude 4.x only, Gemini 2.5 series, current Groq production models
+
+---
+
+## [0.2.0] — 2026-04-03
+
+### Added
+- Autonomous memory system — model emits `<memory>` tags; tags are extracted
+  after each turn and written to the profile's `memories/` directory
+- Three memory categories: `user.md` (stable facts), `notes.md`
+  (role-important), `auto_YYYYMMDD.md` (daily observations)
+- `~/.priests/PRIESTS.md` — memory guide bootstrapped automatically on first
+  run; teaches the model tag syntax and when to save
+- `priests run --no-memories` — disable memory for a single session
+- `memories = false` in `profile.toml` — disable memory permanently for a profile
+- `memories_limit` in `profile.toml` — override global memory file limit
+- `[memory] limit` in `priests.toml` — global cap on daily auto memory files
+- Per-profile `profile.toml` with `memories`, `memories_limit` settings
+- `/new` slash command in interactive chat — starts a new session
+- Arrow key navigation and Ctrl+J newline in interactive chat (prompt_toolkit)
+
+### Fixed
+- Memory tag deduplication — known facts are not re-saved across sessions
+- Placeholder filter — prevents hallucinated `[Unknown]` / `[Name]` values
+  from being written to memory files
+- Session tag cleanup — memory tags stripped from last assistant turn in
+  SQLite so they don't leak into future conversation context
+
+---
+
+## [0.1.0] — 2026-03-31
+
+### Added
+- `priests run` — interactive chat (prompt_toolkit) and single-prompt CLI
+- `priests init` — guided setup: provider selection, API key entry, model pick,
+  writes `~/.priests/priests.toml`
+- `priests config show / set` — read and update config values
+- `priests profile list / init` — manage behavior profiles
+- `priests service start / status` — FastAPI HTTP service on `127.0.0.1:8777`
+- HTTP API: `POST /v1/run`, `POST /v1/chat`, `GET /v1/sessions`,
+  `GET /v1/sessions/{id}`, `GET /health`
+- `/think on` / `/think off` / `/exit` / `/help` slash commands in chat
+- `~/.priests/priests.toml` config with `[default]`, `[paths]`, `[service]`,
+  `[providers.*]`, env var overrides via `PRIESTS_` prefix

priests-0.4.0/DEVLOG.md

@@ -0,0 +1,122 @@
+# DEVLOG
+
+## Future milestones
+
+### Other deferred items
+
+### Other deferred items
+- `service stop` / `service logs` with background daemon mode (PID file)
+- `priests models add` — add/configure additional providers after init
+- `/new` session command in interactive chat
+- Multi-model dispatch, cron task pool, agent orchestration
+- OpenAI provider (tracked in `priest` Milestone 2)
+- `priest` `SqliteSessionStore.list()` method (priests uses direct aiosqlite query as workaround)
+- schedules/plans memory categories (v0.3)
+
+---
+
+## 2026-04-04 — Provider fixes + proxy support
+
+**Root cause: Python 3.14 breaks httpcore's anyio async TLS backend**
+- `AsyncOpenAI` inside `anyio.run()` fails in chat mode (real TTY + prompt_toolkit) due to `httpcore/_backends/anyio.py start_tls` bug on Python 3.14
+- Fix: switched `OpenAICompatProvider` to use sync `OpenAI` client in `anyio.to_thread.run_sync()` — uses httpcore sync sockets instead of anyio async TLS
+- File: `priest/priest/providers/openai_compat_provider.py`
+
+**Proxy support**
+- New `[proxy] url = "..."` section in `priests.toml`
+- `use_proxy = true/false` on each provider config (`OpenAICompatConfig`, `AnthropicConfig`)
+- `engine_factory.py` resolves proxy URL and passes it to adapter constructors
+- `OpenAICompatProvider`, `AnthropicProvider` each accept `proxy: str | None`
+- Proxy passed via `httpx.Client(proxy=...)` / `httpx.AsyncClient(proxy=...)`
+
+**`think` parameter scoped to Qwen/Ollama providers only**
+- Previously sent `{"think": False}` to all providers — breaks Gemini (400 unknown field)
+- `_THINK_PROVIDERS = {"ollama", "bailian", "alibaba_cloud"}` in `run_cmd.py`
+- Only those providers get `think` in `provider_options`; all others get `{}`
+
+**Gemini base URL corrected**
+- Registry had `/openai/v1`, correct is `/v1beta/openai/`
+- Fixed in `registry.py` and updated existing `priests.toml`
+
+---
+
+## 2026-04-03 — Milestone 2: Autonomous memory system
+
+Implemented model-driven persistent memory for `priests` v0.2.
+
+**Memory extraction (`priests/memory/extractor.py`):**
+- Regex-based extraction of `<memory>`, `<memory type="user">`, `<memory type="note">` tags from model responses
+- Three-file routing: `user.md` (stable facts), `notes.md` (role-important), `auto_YYYYMMDD.md` (daily log)
+- Exact-match deduplication (case-insensitive) — no re-saving known facts across sessions
+- Placeholder filter (`[Unknown]`, `[Name]`, etc.) — prevents hallucinated values from being written
+- `trim_memories()` — deletes oldest `auto_YYYYMMDD.md` files beyond configured limit; `user.md` and `notes.md` are never trimmed
+- `clean_last_turn()` — strips memory tags from the last assistant turn in the session store so they don't leak into future context
+
+**Global memory guide (`~/.priests/PRIESTS.md`):**
+- Bootstrapped automatically on first run and on v0.1→v0.2 upgrade
+- Teaches the model tag syntax, format rules, and when to save via a concrete example
+- User-editable; deleting or emptying it disables auto-memory without a code change
+
+**Per-profile config (`priests/profile/config.py`):**
+- `profile.toml` scaffolded by `priests profile init` alongside PROFILE.md/RULES.md
+- `memories = false` — disables memory loading and saving for tool profiles (formatters, dictionaries, etc.)
+- `memories_limit` — overrides the global `[memory].limit` for a specific profile
+- Precedence: CLI flag > `profile.toml` > global `priests.toml`
+
+**CLI (`priests/cli/run_cmd.py`):**
+- `--memories/--no-memories` flag (default on) — disables memory for a single session
+- Chat prompt changed: `you>` → `user >`, `ai>` → `{profile_name} >`
+- `[memory saved]` hint shown after turns where facts were written
+
+**HTTP service (`priests/service/routes/run.py`):**
+- `?memories=false` query param on `/v1/run` and `/v1/chat`
+
+**Config (`priests/config/model.py`):**
+- `MemoryConfig` with `limit: int = 50` — controls how many daily auto files to retain globally
+
+**Verified against:** Ollama + qwen3.5:9b locally. Cross-session recall confirmed (name, preferences, explicit "remember this" requests).
+
+---
+
+## 2026-04-01 — Milestone 1 complete
+
+Initial implementation of `priests` v0.1.0. All planned CLI commands and HTTP service routes are in place and verified against Ollama + qwen3.5:9b.
+
+**Package setup:**
+- `pyproject.toml` with `uv`, hatchling, `priest` as local path dep
+- Added `tool.hatch.metadata.allow-direct-references = true` (required for path deps)
+- Entry point: `priests = "priests.cli.main:app"`
+
+**Config system (`priests/config/`):**
+- `AppConfig` with `DefaultsConfig`, `PathsConfig`, `ServiceConfig`, `ProvidersConfig` (all Pydantic)
+- Config file: `~/.config/priests/priests.toml` (XDG) or `~/.priests/priests.toml`
+- Env var overrides: `PRIESTS_` prefix, `__` for nesting (e.g. `PRIESTS_DEFAULT__MODEL`)
+- `set_config_value()` mutates a dotted key and writes back via `tomli-w`
+- `None` values stripped before TOML serialization (TOML has no null type)
+
+**Engine factory (`priests/engine_factory.py`):**
+- Single function: `build_engine(config) -> (PriestEngine, SqliteSessionStore)`
+- Store returned un-initialized — caller owns lifecycle (`async with store:` for CLI, lifespan for FastAPI)
+
+**CLI (`priests/cli/`):**
+- `priests run [PROMPT]` — single prompt or interactive chat (tty detection); `--no-think` default
+- `priests profile list / init NAME` — scans profiles dir, scaffolds PROFILE.md + RULES.md + memories/
+- `priests config show / set KEY VALUE` — rich TOML display, dotted key mutation
+- `priests service start / status / stop / logs` — foreground uvicorn; status hits `/health`
+
+**HTTP service (`priests/service/`):**
+- FastAPI app via `create_app(config)` with lifespan-managed engine + store
+- `GET /health`, `POST /v1/run`, `POST /v1/chat`, `GET /v1/sessions`, `GET /v1/sessions/{id}`
+- `/v1/chat` auto-generates session ID if none provided
+- Session listing uses direct `aiosqlite` query (SqliteSessionStore has no `list()` — raise PR to priest for v0.2)
+- `SessionSummary` / `SessionDetail` Pydantic models bridge priest's `Session` dataclass to JSON
+
+**Verified against:** Ollama + qwen3.5:9b locally. Session continuity confirmed via `/v1/chat`.
+
+### Out of scope for v0.1.0 (deferred)
+
+- `service stop` / `service logs` (foreground-only; stubs added)
+- Multi-model dispatch, cron task pool, agent orchestration (v1+)
+- OpenAI provider (tracked in `priest` Milestone 2)
+- `priest` `SqliteSessionStore.list()` method (direct aiosqlite query used as workaround)
+- Background service mode with PID file

priests-0.4.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 seesaw game
+
+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.

priests-0.4.0/PKG-INFO

@@ -0,0 +1,17 @@
+Metadata-Version: 2.4
+Name: priests
+Version: 0.4.0
+Summary: CLI tool and HTTP service for priest AI orchestration
+License-File: LICENSE
+Requires-Python: >=3.11
+Requires-Dist: anyio>=4.0
+Requires-Dist: fastapi>=0.111
+Requires-Dist: httpx>=0.27
+Requires-Dist: openai>=2.30.0
+Requires-Dist: priest-core>=0.2.0
+Requires-Dist: pydantic>=2.0
+Requires-Dist: questionary>=2.0
+Requires-Dist: rich>=13.0
+Requires-Dist: tomli-w>=1.0
+Requires-Dist: typer>=0.12
+Requires-Dist: uvicorn[standard]>=0.30