financial-analyst 0.6.0__tar.gz

financial_analyst-0.6.0/.env.example

@@ -0,0 +1,7 @@
+ANTHROPIC_API_KEY=
+OPENAI_API_KEY=
+DASHSCOPE_API_KEY=
+DEEPSEEK_API_KEY=
+TUSHARE_TOKEN=
+FA_LOG_LEVEL=INFO
+FA_CACHE_DIR=~/.financial-analyst/cache

financial_analyst-0.6.0/.gitignore

@@ -0,0 +1,15 @@
+__pycache__/
+*.pyc
+.pytest_cache/
+.ruff_cache/
+*.egg-info/
+build/
+dist/
+.env
+out/
+news/**/*.txt
+news/**/*.html
+news/**/*.pdf
+kb/**/*.parquet
+.venv/
+venv/

financial_analyst-0.6.0/CHANGELOG.md

@@ -0,0 +1,118 @@
+# Changelog
+
+## v0.6.0 — 2026-05-18
+
+### Added
+- First PyPI release. Install: `pip install financial-analyst`.
+- Polished `pyproject.toml` with full metadata (classifiers, urls, keywords, authors).
+
+### Changed
+- README quick-start lead now shows `pip install financial-analyst` instead of `git clone`.
+- Version bump 0.5.0 → 0.6.0.
+
+### Notes
+- No functional code changes vs v0.5.0. This release is packaging-only.
+
+## v0.4.0 — 2026-05-18
+
+### Added (BYOM: Bring Your Own Models)
+- `BaseNewsCollector` ABC — plug-in interface for auto-collecting news into `news/<code>/` drop-zone (`data/collectors/news/base.py`).
+- `BaseF10Collector` ABC — plug-in interface for F10 data (公司公告/龙虎榜/大宗交易) into `f10/<code>/` (`data/collectors/f10/base.py`).
+- 4 example stubs under `examples/`:
+  - `custom_model_fm_cluster.py` — FM cluster model pattern
+  - `custom_loader_csv_only.py` — minimal CSV-backed `BaseLoader`
+  - `custom_news_collector.py` — Tushare news API skeleton
+  - `custom_f10_collector.py` — pytdx F10 skeleton
+- Plugin discovery: `config/plugins.yaml` lists user `.py` files exec'd at startup (`src/financial_analyst/plugins.py`).
+- CLI introspection: `financial-analyst {models,loaders,agents,collectors} list`.
+- `docs/byom.md` — full Bring-Your-Own-Models guide.
+
+### Changed
+- README "Extending" section now points to BYOM workflow.
+- `tests/test_agent_registry.py` no longer pollutes the `SubAgentRegistry`; fixtures clear it.
+
+## v0.3.0 — 2026-05-18
+
+### Added (Ingest + Dream Loop)
+- **CSV → Qlib binary ingester** (`data/ingest/csv_ingester.py`) with both long-format and per-code-filename support, schema-configurable, ohlcv field mapping.
+- `BaseIngester` ABC + reserved `AkshareIngester` / `YfinanceIngester` stubs for v0.4+.
+- CLI: `financial-analyst ingest --source <name> [--dry-run]`.
+- **Dream loop** for agent self-improving memory:
+  - `OutcomeTracker` — measure T+5d/T+20d outcomes against past predictions in `out/*.json`, scoring verdict ∈ {correct, wrong, partial, pending}.
+  - `Introspector` sub-agent — LLM-driven post-mortem analyst (NOT in stock-deep-dive preset).
+  - `ProposalWriter` — writes `Introspector` proposals to `memories/_proposed/<agent>/<date>_<slug>.md` with YAML frontmatter.
+  - `memories/introspector/introspector_rules.md` meta-rules (focus on wrong>partial>correct, 2/3-5/6+ confidence thresholds, target risk-officer when in doubt).
+- CLI: `financial-analyst dream [--since 30] [--dry-run]`.
+- TUI: `/dream`, `/memory list-proposals`, `/memory accept _proposed/<file>`, `/memory reject _proposed/<file>`.
+- `docs/data_ingest.md`, `docs/dream_loop.md`.
+
+### Changed
+- `config/data_sources.yaml` template added for ingester config.
+- Memory CLI usage strings updated to list all 11 subcommands.
+
+### Safety
+- Dream proposals require human review (no auto-merge); auto-accept is explicitly NOT implemented.
+- `/memory accept` only operates on paths starting with `_proposed/`.
+
+## v0.2.3 — 2026-05-18
+
+### Fixed (Hotfix found during real SH600666 testing)
+- **`AgentMemory.load_relevant` falls back to `load_all` on 0 FTS5 hits** — prevents agents going "blind" when the JSON-derived query doesn't match memory wording.
+- **Per-agent `always_include.txt`** — listed files load unconditionally regardless of retrieval results. Initial entry: `memories/risk-officer/always_include.txt` lists `hard_rules.md` (game-capital veto must never be missed).
+- **`report-writer` post-validation** — if `risk-officer.veto_flags` is non-empty OR `rating_overall ≤ 0`, `position_pct` is forced to 0 and `action` re-derived. Sanity-override notes appended to the markdown report.
+- **`mv_tier` enum** — `fundamental-analyst.FundamentalOutput.mv_tier` changed from `str` to `Literal["large","mid","small"]`; pre-normalize Chinese variants (`中小盘`→`small`, `大盘`→`large`, etc.) before pydantic validation.
+
+## v0.2.2 — 2026-05-18
+
+### Added
+- **5min bar support**. `QlibBinaryLoader` now accepts `dict` provider_uri with `day` + `5min` (+ optional `1min`) keys.
+- `BaseLoader.fetch_quote` signature extended with `freq: str = "day"` (backward compatible).
+- `factor-computer` auto-fetches 5min bars where available, activating:
+  - **board_scorer v5 `seal_micro` dimension** (-3..+3): `seal_bar`, `seal_at_close`, `gap_open`, `open_count`.
+  - **volume_regime R11 `tail_surge`** signal: last-30-min volume + return ramp.
+  - **R14 super_distr** combined signal (`r9_distr AND r11_tail_surge`).
+- TushareLoader gracefully returns empty DataFrame for non-day freq.
+
+## v0.2.1 — 2026-05-18
+
+### Added
+- `ParquetCache` wired into `TushareLoader` (cache miss → API; cache hit → no network). Configurable TTL (default 86400s = 1 day) and `enable_cache=False` opt-out.
+- `QlibBinaryLoader` reads Qlib binary directories — zero-network microsecond reads. Schema: `<provider_uri>/calendars/day.txt` + `instruments/all.txt` + `features/<code_lower>/<field>.day.bin` (4-byte float32 start_index + float32 array).
+- `loader_factory.get_default_loader()` — reads `config/loaders.yaml` to instantiate the configured default. Sub-agents (`quote-fetcher`, `factor-computer`, `LGBMomentumModel`) use this factory.
+- `config/loaders.yaml` template with both `tushare` (cache) and `qlib_binary` options.
+
+### Changed
+- `TushareLoader` re-implemented using raw `requests.post` (bypasses the `tushare` Python library's round-robin to `api.waditu.com` which times out behind corporate proxies). HTTP only.
+- `quote-fetcher` uses `_safe_float` for `daily_basic` fields (handles None/NaN gracefully for stocks without dividends, etc.).
+- `cli.py` calls `load_dotenv(override=True)` so `.env` overrides shell env vars (fixes Windows user-level TUSHARE_TOKEN conflicts).
+- `llm.client` now explicitly passes `api_key` from per-provider `api_key_env` so LiteLLM doesn't fall back to OPENAI_API_KEY when using qwen/deepseek/etc.
+- `config/llm.yaml` default switched to Qwen (`qwen3.5-plus`) since most users have DashScope keys, not Anthropic.
+- Report writer renders the markdown report inline in terminal (Rich Markdown) and exports a colored HTML copy next to the .md. Clickable `file:///` URL in TUI output.
+- Forced UTF-8 stdout/stderr at module load (cli.py + tui.py) so Windows zh-CN PowerShell doesn't choke on `¥` / emoji / rare CJK.
+
+## v0.2.0 — 2026-05-18
+
+### Added
+- `MemoryIndex` — SQLite FTS5 full-text index over `memories/**/*.md` with CJK tokenization, incremental updates, agent-filtered search.
+- `AgentMemory.load_relevant(query, top_k)` — hybrid retrieval that pulls top-K snippets via FTS5 while always including `_shared/` core rules. Backward-compatible with `load_all()`.
+- Per-agent `memory_mode: full | retrieval` configuration in swarm preset YAML. Defaults to `full` (v0.1 behavior preserved).
+- TUI `/memory` subcommands: `search`, `show`, `edit`, `stats`, `diff`, `reindex` (in addition to existing `list`, `reload`).
+- `bear-advocate` and `risk-officer` opted into retrieval mode by default (biggest memory libraries).
+
+### Changed
+- `SubAgent.__init__` accepts optional `index: Optional[MemoryIndex] = None`.
+- `swarm.load_preset()` accepts `memory_index` parameter; passes through to retrieval-mode agents only.
+- `MemoryIndex.stats()` now includes `total_bytes` and `per_agent_bytes`.
+
+### Token cost impact
+- Single-stock report: ~80K → ~30K tokens (estimated 62% reduction) when both retrieval-mode agents are exercised.
+- Per-report Qwen cost: ~¥0.05 → ~¥0.02.
+
+## v0.1.0 — 2026-05-17
+
+### Initial release
+- 13 sub-agents in 3 trust tiers (5 Tier-1 data fetchers, 4 Tier-2 analysts, 4 Tier-3 decision agents).
+- Pluggable per-agent memory (`memories/<agent>/*.md`) with `_shared/` cross-agent playbook.
+- Tushare data loader, LGB momentum model, LiteLLM multi-provider abstraction.
+- Rich TUI with prompt-toolkit REPL.
+- 100+ pydantic-validated tests, opt-in real E2E test.

financial_analyst-0.6.0/CLAUDE.md

@@ -0,0 +1,18 @@
+# Claude Code Working Instructions
+
+Multi-agent A-share research workstation. 13 sub-agents in three trust tiers.
+
+## Hard Rules
+- Only `report-writer` has the `write` tool. All others are read-only.
+- Untrusted sources (news, F10) only touch `news-reader` and `f10-reader`. Their output is pydantic-validated JSON.
+- Memory is per-agent under `memories/<agent>/*.md`. Hot-reloadable: edit a markdown → next agent invocation picks it up.
+- LLM is provider-abstracted via LiteLLM. Configure in `config/llm.yaml`.
+
+## Code Style
+- Python 3.11+, type hints, async-first (asyncio.gather for parallel tiers).
+- pydantic v2 for all sub-agent IO.
+- pytest for tests. Mock LLM in unit tests; real LLM only in integration tests under `tests/integration/`.
+
+## Workflow
+- Each task in `docs/superpowers/plans/*.md` should produce exactly one commit.
+- Run `pytest tests/` before each commit.

financial_analyst-0.6.0/LICENSE

@@ -0,0 +1,195 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship made available under
+      the License, as indicated by a copyright notice that is included in
+      or attached to the work (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other transformations
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and its derivative works thereof.
+
+      "Contribution" shall mean, as submitted to the Licensor for inclusion
+      in the Work by the copyright owner or by an individual or Legal Entity
+      authorized to submit on behalf of the copyright owner. For the purposes
+      of this definition, "submitted" means any form of electronic, verbal,
+      or written communication sent to the Licensor or its representatives,
+      including but not limited to communication on electronic mailing lists,
+      source code control systems, and issue tracking systems that are managed
+      by, or on behalf of, the Licensor for the purpose of tracking and
+      discussing modifications to the Work, but excluding communication that
+      is conspicuously marked or designated in writing by the copyright owner
+      as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any Legal Entity on behalf of
+      whom a Contribution has been received by the Licensor and included
+      within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by the combined work (in which such
+      Contribution(s) was submitted). If You institute patent litigation
+      against any entity (including a cross-claim or counterclaim in a
+      lawsuit) alleging that the Work or any combination of the Work
+      constitutes direct or contributory patent infringement, then any
+      patent licenses granted to You under this License for that Work
+      shall terminate as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or Derivative
+          Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, You must include a readable copy of the
+          attribution notices contained within such NOTICE file, in
+          at least one of the following places: within a NOTICE text
+          file distributed as part of the Derivative Works; within
+          the Source form or documentation, if provided along with the
+          Derivative Works; or, within a display generated by the
+          Derivative Works, if and wherever such third-party notices
+          normally appear. The contents of the NOTICE file are for
+          informational purposes only and do not modify the License.
+          You may add Your own attribution notices within Derivative
+          Works that You distribute, alongside or in addition to the
+          NOTICE text from the Work, provided that such additional
+          attribution notices cannot be construed as modifying the
+          License.
+
+      You may add Your own license statement for Your modifications and
+      may provide additional grant of rights to use, copy, modify, merge,
+      publish, distribute, sublicense, and/or sell copies of the Work,
+      and to permit persons to whom the Work is furnished to do so,
+      subject to the following conditions: you must include the above
+      copyright notice and this permission notice in all copies or
+      substantial portions of the Work.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or reproducing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or exemplary damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or all other
+      commercial damages or losses), even if such Contributor has been
+      advised of the possibility of such damages.
+
+   9. Accepting Warranty or Liability. While redistributing the Work or
+      Derivative Works thereof, You may choose to offer, and charge a fee
+      for, acceptance of support, warranty, indemnity, or other liability
+      obligations and/or rights consistent with this License. However, in
+      accepting such obligations, You may offer such obligations only on
+      Your own behalf and on Your sole responsibility, not on behalf of
+      any other Contributor, and only if You agree to indemnify, defend,
+      and hold each Contributor harmless for any liability incurred by,
+      or claims asserted against, such Contributor by reason of your
+      accepting any warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format in question. It may also be
+      a square bracket around a slash like "[/]" for display.
+
+      Copyright 2026 financial-analyst contributors
+
+      Licensed under the Apache License, Version 2.0 (the "License");
+      you may not use this file except in compliance with the License.
+      You may obtain a copy of the License at
+
+          http://www.apache.org/licenses/LICENSE-2.0
+
+      Unless required by applicable law or agreed to in writing, software
+      distributed under the License is distributed on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied. See the License for the specific language governing
+      permissions and limitations under the License.

financial_analyst-0.6.0/PKG-INFO

@@ -0,0 +1,217 @@
+Metadata-Version: 2.4
+Name: financial-analyst
+Version: 0.6.0
+Summary: A-share single-stock deep-dive multi-agent research workstation with three-tier trust isolation, FTS5 memory retrieval, dream loop introspection, and BYOM extensibility.
+Project-URL: Homepage, https://github.com/jesson-hh/financial-analyst
+Project-URL: Repository, https://github.com/jesson-hh/financial-analyst
+Project-URL: Issues, https://github.com/jesson-hh/financial-analyst/issues
+Project-URL: Changelog, https://github.com/jesson-hh/financial-analyst/blob/main/CHANGELOG.md
+Project-URL: Documentation, https://github.com/jesson-hh/financial-analyst/tree/main/docs
+Author-email: jesson-hh <xuyi1030@proton.me>
+Maintainer-email: jesson-hh <xuyi1030@proton.me>
+License: Apache-2.0
+License-File: LICENSE
+Keywords: a-share,ai-agent,claude,fintech,fts5,investment-research,litellm,llm,multi-agent,qlib,quantitative-finance,qwen,stock-research,tushare
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Natural Language :: Chinese (Simplified)
+Classifier: Natural Language :: English
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Office/Business :: Financial :: Investment
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.11
+Requires-Dist: anyio>=4
+Requires-Dist: lightgbm>=4
+Requires-Dist: litellm>=1.50
+Requires-Dist: numpy>=1.26
+Requires-Dist: pandas>=2
+Requires-Dist: prompt-toolkit>=3.0
+Requires-Dist: pyarrow>=15
+Requires-Dist: pydantic-settings>=2.1
+Requires-Dist: pydantic>=2.5
+Requires-Dist: pyyaml>=6
+Requires-Dist: rich>=13
+Requires-Dist: tushare>=1.4
+Requires-Dist: typer>=0.12
+Provides-Extra: dev
+Requires-Dist: build>=1.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
+Requires-Dist: pytest-mock>=3.12; extra == 'dev'
+Requires-Dist: pytest>=8; extra == 'dev'
+Requires-Dist: ruff>=0.5; extra == 'dev'
+Requires-Dist: twine>=5.0; extra == 'dev'
+Description-Content-Type: text/markdown
+
+# Financial Analyst
+
+[![PyPI](https://img.shields.io/pypi/v/financial-analyst)](https://pypi.org/project/financial-analyst/)
+[![tests](https://img.shields.io/badge/tests-240_passed-brightgreen)](https://github.com/jesson-hh/financial-analyst/actions)
+[![python](https://img.shields.io/badge/python-3.11+-blue)](https://www.python.org/)
+[![license](https://img.shields.io/badge/license-Apache_2.0-green)](LICENSE)
+
+**A-share single-stock deep-dive multi-agent research workstation.**
+
+**13 sub-agents in three trust tiers** — five Tier-1 data fetchers (two of which read untrusted news/F10 with JSON-schema-locked output), four Tier-2 analysts (fundamental, technical, whale-sentiment, quant), four Tier-3 decision agents (bull, bear, risk officer, report writer). Only the report writer can write files. Memory is pluggable per-agent (`memories/<agent>/*.md`) — edit a markdown, the next report uses it. FTS5-backed retrieval keeps prompt costs ~60% lower than naive full-injection.
+
+Inspired by [Anthropic's financial-services](https://github.com/anthropics/financial-services) (3-tier trust isolation, single-writer pattern) and [HKUDS/Vibe-Trading](https://github.com/HKUDS/Vibe-Trading) (YAML swarm presets, multi-provider LLM).
+
+## What's in v0.6
+
+- **13 sub-agent DAG** (data → analyst → decision), Qwen/Claude/DeepSeek/Ollama via LiteLLM
+- **QlibBinaryLoader** (day + 5min) + **TushareLoader** (HTTP + ParquetCache) + **CSV ingester**
+- **R7-R20 sentiment signals** (board_scorer v5, volume_regime super_distr/tail_surge, whale signals)
+- **Pluggable memory** with FTS5 retrieval + always_include white-list + `_shared/` cross-agent playbook
+- **Dream loop** — agent self-improving memory (OutcomeTracker + Introspector + `memories/_proposed/` staging + human accept/reject)
+- **BYOM** — registry-based plug-in points for models / loaders / collectors / sub-agents / KBs; `config/plugins.yaml` auto-loads user `.py` files at startup
+- **Markdown + JSON + HTML** output (Rich Markdown rendering + standalone HTML file)
+- **226 tests** unit + integration (mocked LLM)
+
+## Quick Start
+
+**One-line install (when v0.6.0 is on PyPI):**
+
+```bash
+pip install financial-analyst
+```
+
+Or **from source** (development):
+
+```bash
+git clone https://github.com/jesson-hh/financial-analyst.git
+cd financial-analyst
+python -m venv .venv && .venv\Scripts\activate    # Windows; or `source .venv/bin/activate`
+pip install -e .[dev]
+```
+
+Then:
+
+```bash
+cp .env.example .env
+# Edit .env: set TUSHARE_TOKEN + your LLM provider key (default config uses DASHSCOPE_API_KEY for Qwen)
+financial-analyst                                  # boots TUI
+```
+
+In TUI:
+
+```
+> 看看 600519
+> /ask 我最近研报里 bear 段是不是过激了
+> /agents
+> /memory search 游资
+> /show
+> /dream --since 30
+> /memory list-proposals
+> /quit
+```
+
+One-shot:
+
+```bash
+financial-analyst report SH600519
+financial-analyst ask "SH600519 现在 PE 多少"
+financial-analyst ingest --source my_csv
+financial-analyst dream --since 30
+financial-analyst models list
+```
+
+## Architecture
+
+```
+Orchestrator → Tier 1 (data, parallel) → Tier 2 (analysts, parallel) → Tier 3 (decision, serial)
+                  ↓                          ↓                              ↓
+              fetch + factor              4 analysts                  bull/bear/risk → writer
+              + read untrusted            consume tier 1 JSON         consume tier 2 JSON
+              with JSON schemas
+```
+
+See [docs/architecture.md](docs/architecture.md) for the full DAG and trust model.
+
+## Extending — Bring Your Own Models (BYOM)
+
+`financial-analyst` is a **framework**, not a fixed product. Plug in your own private models / loaders / collectors:
+
+```python
+# G:/my_private_code/my_fm.py
+from financial_analyst.models import BaseModel, ModelRegistry
+
+class MyFMCluster(BaseModel):
+    def predict(self, code, asof):
+        return {"score": ..., "rank_pct": ..., "cluster": ...}
+    def metadata(self):
+        return {"name": "my_fm", "version": "W10"}
+
+ModelRegistry.register("my_fm", MyFMCluster)
+```
+
+```yaml
+# config/plugins.yaml
+load_at_startup:
+  - G:/my_private_code/my_fm.py
+```
+
+Now `financial-analyst report SH600519` includes your model in the quant consensus. **No proprietary checkpoint enters the open-source repo.**
+
+See [docs/byom.md](docs/byom.md) for the full guide. Examples for FM cluster / CSV loader / Tushare news collector / pytdx F10 collector are in [`examples/`](examples/).
+
+## Data Ingestion
+
+If you don't have a Qlib data directory, ingest your CSVs:
+
+```yaml
+# config/data_sources.yaml
+sources:
+  - name: my_csv
+    type: csv
+    path: G:/my_data/*.csv
+    code_col: ts_code
+    date_col: trade_date
+    target: ~/.financial-analyst/data/my_csv
+```
+
+```bash
+financial-analyst ingest --source my_csv
+# Then point config/loaders.yaml `qlib_binary.provider_uri.day` at the target
+```
+
+See [docs/data_ingest.md](docs/data_ingest.md).
+
+## Memory System
+
+Each sub-agent has `memories/<agent-name>/` with markdown files. Files are appended to the agent's system prompt at runtime. v0.2+: FTS5 retrieval keeps prompt cost down for agents marked `memory_mode: retrieval`. Critical files listed in `memories/<agent>/always_include.txt` are loaded unconditionally.
+
+See [docs/memories.md](docs/memories.md) for principles, file-organization advice, and CLI commands.
+
+## Dream Loop — Agent Self-Improving Memory
+
+Run `/dream` (or `financial-analyst dream`) to:
+1. Score past reports against T+5d / T+20d actual prices.
+2. Have an introspector sub-agent identify patterns in wrong predictions.
+3. Stage proposals in `memories/_proposed/<agent>/` for human review.
+4. `/memory accept` merges; `/memory reject` discards.
+
+**Auto-accept is intentionally NOT implemented** — incorrect memory updates compound losses in quant systems. Human review only.
+
+See [docs/dream_loop.md](docs/dream_loop.md).
+
+## Tests
+
+```bash
+pytest tests/                                       # 240 unit + integration tests (mocked)
+FA_E2E=1 pytest tests/integration/test_end_to_end.py  # real Tushare + LLM round-trip
+```
+
+## License
+
+Apache 2.0.
+
+## Disclaimer
+
+Drafts analyst work product for review by qualified professionals. Does not make investment recommendations, execute transactions, or post to any ledger. You are responsible for compliance with applicable laws and regulations.