backtestchat 0.1.0__tar.gz

backtestchat-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,35 @@
+name: Publish to PyPI
+
+# Publishes backtestchat to PyPI when you cut a GitHub Release.
+# Auth is PyPI Trusted Publishing (OIDC) - no API tokens stored anywhere.
+# The publish is gated on the test suite passing.
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/backtestchat
+    permissions:
+      id-token: write   # required for OIDC trusted publishing
+      contents: read
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v5
+
+      - name: Run tests (gate the publish)
+        run: |
+          uv sync --group dev
+          uv run pytest -q
+
+      - name: Build sdist + wheel
+        run: uv build
+
+      - name: Publish to PyPI (trusted publishing)
+        uses: pypa/gh-action-pypi-publish@release/v1

backtestchat-0.1.0/.gitignore

@@ -0,0 +1,35 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+build/
+dist/
+
+# Virtual environments
+.venv/
+venv/
+env/
+
+# Environment / secrets
+.env
+.env.*
+
+# Runtime data (created on startup)
+data/
+artifacts/
+*.sqlite3
+*.sqlite3-journal
+*.sqlite3-wal
+*.sqlite3-shm
+
+# Throwaway manual smoke scripts (not part of the package)
+scripts/
+
+# Tooling
+.pytest_cache/
+.ruff_cache/
+.mypy_cache/
+.vscode/
+
+# uv.lock is committed on purpose - do not ignore it.

backtestchat-0.1.0/AGENTS.md

@@ -0,0 +1,42 @@
+# AGENTS.md
+
+Guidance for coding agents and contributors working in this repository. This
+mirrors the key rules in CLAUDE.md; see CLAUDE.md for stack, domain decisions,
+and layout.
+
+## What this project is
+
+backtestchat: an open-source, local-first agentic strategy lab CLI. A LangGraph
+tool-calling agent fetches keyless Binance public market data, backtests
+long/flat strategies, and runs simulated paper trades. Paper-only by design -
+no exchange keys, no real orders.
+
+## House rules (non-negotiable)
+
+- Plain ASCII everywhere (copy, comments, docs). No em/en dashes, curly quotes,
+  or ellipsis characters. Use `-` and `...` instead.
+- Loud, structured, self-repairing tool errors:
+  `{"error": code, "message": ..., "valid_options": [...]}`. No fuzzy parsing
+  inside tools; canonical encodings only.
+- Data tools never render; the LLM never writes HTML. Rendering is a
+  deterministic function of widget dicts.
+- Pure logic stays pure (engine/signals: bars in, results out; no I/O, globals).
+- Idempotence over scheduling: the processed-bar-ts guard is the cadence
+  control; any command may safely trigger a sync at any time.
+
+## Scope
+
+Paper-only, keyless data, no daemon, and exactly 10 agent tools. See the "Scope
+and roadmap" section of CLAUDE.md for what is intentionally out of scope and
+should be proposed as an addition rather than assumed to belong in the core.
+
+## Build and test
+
+```
+uv sync
+uv run pytest -q     # deterministic, no network, no model key
+```
+
+Package manager is uv; never edit uv.lock by hand (let `uv` resolve it). Tests
+must stay offline: mock HTTP for the market layer and use a scripted fake model
+for the agent graph.

backtestchat-0.1.0/CLAUDE.md

@@ -0,0 +1,107 @@
+# backtestchat - contributor and coding-agent guide
+
+backtestchat is an open-source, local-first agentic strategy lab CLI: a LangGraph
+tool-calling agent for fetching keyless Binance public market data, backtesting
+long/flat strategies, and running simulated paper trades. Paper-only by design.
+
+This file captures the house rules and design decisions so changes stay
+consistent. If a rule here conflicts with a maintainer's instruction, follow the
+maintainer.
+
+## Design principles (house rules, apply everywhere)
+
+- Fix tool surfaces: canonical encodings, alias tables where needed, loud
+  structured errors that teach the model the valid options
+  (`{"error": code, "message": ..., "valid_options": [...]}`). Never add
+  inference layers, fuzzy parsing, or special-cased conditions inside tools; the
+  model maps user language to canonical values.
+- Data tools do not render. Rendering is a deterministic function of widget
+  dicts. The LLM never writes HTML or markup.
+- Pure logic stays pure: engine and signals take bars in, return results out -
+  no I/O, no globals. This is what makes them testable and portable.
+- Idempotence over scheduling: the processed-bar-ts guard is the cadence
+  control; any command may safely trigger a sync at any time.
+- Plain ASCII in all copy, comments, and docs. No em/en dashes, no curly quotes,
+  no ellipsis characters. Use `-` and `...` instead.
+
+## Stack
+
+- Python >= 3.12, uv + pyproject.toml, src/ layout, console script `backtestchat`.
+- langgraph (agent graph + SqliteSaver checkpointer). langchain-core (@tool,
+  messages) + init_chat_model for BYOK. No legacy LangChain chains/agents.
+- typer + rich (CLI + streaming output). httpx (Binance client; no
+  python-binance). plotly (chart HTML via fig.write_html).
+- sqlite3 via a thin hand-rolled store (no SQLAlchemy, no migrations;
+  CREATE TABLE IF NOT EXISTS on startup). LangGraph checkpoints live in their
+  own sqlite file via SqliteSaver.
+- pytest; engine and paper-run logic tested with fixture bars, deterministic,
+  no network.
+
+No Celery, Redis, MySQL, FastAPI, or frontend framework. Single process, single
+user, local files.
+
+## Domain decisions
+
+- Symbols: Binance spot (BTCUSDT, ETHUSDT, SOLUSDT, ...). Validated against a
+  cached GET /api/v3/exchangeInfo symbol set; unknown symbols fail loudly with
+  suggestions.
+- Intervals (Binance strings, canonical): 15m, 1h, 4h, 1d. bars_per_year:
+  15m=35040, 1h=8760, 4h=2190, 1d=365 (crypto trades 24/7/365).
+- The still-forming bar is always dropped: only closed bars enter the engine,
+  paper runner, or charts.
+- Bar contract everywhere: `{ts, open, high, low, close, volume}` with ts an ISO
+  UTC string.
+- Binance client: public keyless, base https://api.binance.com. klines
+  (limit 1000, paginate by startTime), ticker for quotes, exchangeInfo for
+  validation (cached 24h). In-process TTL caches (quote 30s, klines 60s). Retry
+  twice with backoff; HTTP errors surfaced verbatim.
+- Engine: long/flat only, next-bar fill, cost_bps wired but default 0. Metrics:
+  CAGR, Sharpe, max drawdown, win rate, trade count.
+- Paper runs: the first fill is on the first bar that CLOSES after creation (no
+  backdated fills). Advancing replays all closed bars since the last processed
+  bar, one signal eval per bar; the already-processed-ts guard is the only
+  cadence control. Fills are fractional and fully allocated. Default starting
+  cash 10000 USDT.
+
+## Layout
+
+```
+src/backtestchat/
+  cli.py        typer app: chat, runs, show, config
+  config.py     env > ./backtestchat.toml > defaults
+  artifacts.py  rendered-artifact resolution (last/newest/path)
+  agent/        graph.py, tools.py, prompt.py, context.py, chat.py
+  broker/       base.py (interface, window logic), binance.py (httpx client)
+  strategy/     registry.py, signals.py, indicators.py, engine.py
+  paper/        runner.py (replay), store.py (sqlite)
+  render/       widgets.py (dict builders), plotly_render.py (HTML)
+tests/          fixtures/, test_engine.py, test_registry.py, ...
+```
+
+`./data` (sqlite store + LangGraph checkpoints) and `./artifacts` (chart HTML)
+are created at runtime and gitignored.
+
+## Scope and roadmap
+
+Intentionally in scope: paper-only simulation, keyless data, the 10 agent tools
+(get_quote, get_price_series, list_strategies, run_backtest, compare_strategies,
+save_strategy, list_saved_strategies, start_paper_run, list_paper_runs,
+stop_paper_run), and the ported strategy catalog.
+
+Deliberately out of scope (a clean adapter/extension, not yet built): Binance
+testnet order routing (the Broker interface is designed to make this an adapter
+add), a news-signal tool, web/FastAPI/TUI charts, live streaming / alerts /
+daemons, portfolio / multi-asset / shorting / position sizing / leverage, and
+new strategies beyond the catalog. Propose these as additions rather than
+assuming they belong in the core.
+
+## Build and test
+
+```
+uv sync
+uv run pytest -q     # deterministic, no network, no model key
+```
+
+The engine, signals, broker window logic, and paper runner are pure and tested
+against fixture bars; the market-data and agent layers use mocked HTTP and a
+scripted fake model, so the whole suite runs offline.

backtestchat-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Baris Arat
+
+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.

backtestchat-0.1.0/PKG-INFO

@@ -0,0 +1,198 @@
+Metadata-Version: 2.4
+Name: backtestchat
+Version: 0.1.0
+Summary: Local-first agentic strategy lab CLI: chat, backtest, and paper-trade long/flat strategies on keyless Binance public data.
+Project-URL: Homepage, https://github.com/barisarat/backtestchat
+Project-URL: Repository, https://github.com/barisarat/backtestchat
+Project-URL: Issues, https://github.com/barisarat/backtestchat/issues
+Author: Baris Arat
+License: MIT
+License-File: LICENSE
+Keywords: agent,backtest,binance,cli,langgraph,paper-trading,trading
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Office/Business :: Financial :: Investment
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.12
+Requires-Dist: httpx>=0.27.0
+Requires-Dist: langchain-core>=0.3.0
+Requires-Dist: langchain>=0.3.0
+Requires-Dist: langgraph-checkpoint-sqlite>=2.0.0
+Requires-Dist: langgraph>=0.2.60
+Requires-Dist: plotly>=5.22.0
+Requires-Dist: rich>=13.7.0
+Requires-Dist: typer>=0.12.0
+Provides-Extra: anthropic
+Requires-Dist: langchain-anthropic>=0.2.0; extra == 'anthropic'
+Provides-Extra: ollama
+Requires-Dist: langchain-ollama>=0.2.0; extra == 'ollama'
+Provides-Extra: openai
+Requires-Dist: langchain-openai>=0.2.0; extra == 'openai'
+Description-Content-Type: text/markdown
+
+# backtestchat
+
+An open-source, local-first agentic strategy lab in your terminal. Chat with a
+LangGraph agent to fetch keyless Binance public market data, design and backtest
+long/flat crypto strategies, and run simulated paper trades. Charts render to
+self-contained plotly HTML that opens in your browser.
+
+**Paper-only by design.** backtestchat simulates fills from closed bars. It never
+places real orders and never needs exchange API keys. Market data comes from the
+public, keyless Binance REST API. The only key you supply is your own model
+(BYOK) key, or a local Ollama endpoint.
+
+> backtestchat is a research and simulation tool and a portfolio/learning project.
+> It is not financial advice and not a trading product. Do not use it to make
+> investment decisions.
+
+<!-- TODO: asciinema GIF of a chat session (quote -> backtest -> chart). -->
+<!-- TODO: screenshot of a backtestReport chart. -->
+
+## Quickstart
+
+Requires Python 3.12+ and [uv](https://docs.astral.sh/uv/).
+
+```bash
+git clone https://github.com/barisarat/backtestchat
+cd backtestchat
+uv sync --extra openai              # deps + the OpenAI model provider
+
+export BACKTESTCHAT_MODEL=openai:gpt-5.4-nano
+export OPENAI_API_KEY=sk-...
+
+uv run backtestchat chat                # talk to the agent
+uv run backtestchat runs sync           # advance paper runs to the latest closed bar
+uv run backtestchat runs list           # check paper run status
+```
+
+Two things to know:
+
+- The `backtestchat` command lives in the project venv, so either prefix every
+  command with `uv run`, or activate the venv once with
+  `source .venv/bin/activate` and call `backtestchat` directly.
+- Always sync with the extra for the provider you use. A plain `uv sync`
+  removes extras, and `backtestchat chat` will fail until you
+  `uv sync --extra openai` again.
+
+Market data works with **zero keys**; only the chat agent needs a model key.
+Other providers (Anthropic, Ollama, ...) are covered under
+[Model setup](#model-setup-byok).
+
+## What it does
+
+- **Chat** with a tool-calling agent (LangGraph) that maps your plain-English
+  requests onto market-data, backtest, and paper-trade actions.
+- **Backtest** ~11 single-signal long/flat strategies (SMA/EMA crossover, MACD,
+  RSI, Bollinger, Donchian, ROC, stochastic, ATR channel, price-vs-SMA,
+  buy & hold) with CAGR, Sharpe, max drawdown, win rate, and trade count.
+- **Compare** several strategies at once, ranked, on one chart.
+- **Paper-trade** a strategy with simulated fills. Runs advance by replaying
+  closed bars - no daemon, no scheduler. Any command can trigger a cheap sync.
+- **Charts** are deterministic plotly HTML written to `./artifacts`. The model
+  never writes HTML; it only triggers the tools that build the charts.
+
+Supported intervals: **15m, 1h, 4h, 1d**. Symbols: any Binance spot pair
+(BTCUSDT, ETHUSDT, SOLUSDT, XRPUSDT, DOGEUSDT, ...).
+
+## Model setup (BYOK)
+
+backtestchat talks to any provider via LangChain's `init_chat_model`, configured with
+a `provider:model` string in `BACKTESTCHAT_MODEL` (or `model = "..."` in
+`./backtestchat.toml`). The default setup uses OpenAI:
+
+```bash
+uv sync --extra openai
+export BACKTESTCHAT_MODEL=openai:gpt-5.4-nano
+export OPENAI_API_KEY=sk-...
+```
+
+Any provider `init_chat_model` supports works the same way: install the matching
+extra (`--extra anthropic`, `--extra ollama`, ...) and set `BACKTESTCHAT_MODEL` to
+that provider's `provider:model` string plus its API key (Ollama runs locally
+and needs no key).
+
+If no model is configured, `backtestchat chat` prints these instructions and exits.
+Check what is resolved with `backtestchat config`.
+
+## Usage
+
+```bash
+uv run backtestchat chat
+```
+
+Then talk to it naturally:
+
+```
+you > what's BTC trading at?
+you > backtest an ma crossover 20/50 on BTCUSDT 1h over the last 3 months
+you > compare that against buy and hold and rsi
+you > save that strategy            # asks you to confirm in the terminal
+you > start a paper run with 5000 usdt   # asks you to confirm
+you > list my paper runs
+```
+
+Charts open automatically in your browser and are saved under `./artifacts`.
+Saving a strategy and starting a paper run pause for a y/n confirmation
+(human-in-the-loop, implemented with LangGraph interrupts).
+
+### Commands
+
+```bash
+backtestchat chat [--new] [--thread ID]   # agent REPL (resumes the last thread by default)
+backtestchat runs list [--status active]  # list paper runs (stored state)
+backtestchat runs sync                    # advance all active runs to the latest closed bar
+backtestchat show [last|FILE]             # open a rendered chart artifact
+backtestchat config                       # print resolved configuration
+```
+
+### How paper runs work
+
+A paper run does not backdate: the first simulated fill happens on the first bar
+that **closes after** you create the run, so a fresh run shows
+"waiting for first tick" until then. Advancing a run replays every closed bar
+since its last processed bar, one signal evaluation per bar. Because the
+already-processed-bar guard is the only cadence control, `backtestchat runs sync` is
+safe to run anytime and is idempotent.
+
+## LangSmith tracing (optional)
+
+Set the standard env vars and traces appear in your LangSmith project - no code
+or flags required:
+
+```bash
+export LANGSMITH_TRACING=true
+export LANGSMITH_API_KEY=...
+```
+
+## Configuration
+
+Resolved in order: environment variables > `./backtestchat.toml` > defaults.
+
+| Key | Env var | Default |
+|---|---|---|
+| model | `BACKTESTCHAT_MODEL` | (unset; required for chat) |
+| data dir | `BACKTESTCHAT_DATA_DIR` | `./data` |
+| artifacts dir | `BACKTESTCHAT_ARTIFACTS_DIR` | `./artifacts` |
+
+`./data` holds the sqlite store (strategies, paper runs, trades) and the
+LangGraph checkpoint database. `./artifacts` holds rendered chart HTML. Both are
+gitignored.
+
+## Development
+
+```bash
+uv sync --extra openai  # dev dependencies included; keep your provider extra
+uv run pytest -q        # deterministic, no network, no model key
+```
+
+The engine, signals, broker window logic, and paper runner are pure and tested
+against fixture bars. Market-data and agent layers are tested with mocked HTTP
+and a scripted fake model, so the whole suite runs offline.
+
+## License
+
+MIT - see [LICENSE](LICENSE).

backtestchat-0.1.0/README.md

@@ -0,0 +1,163 @@
+# backtestchat
+
+An open-source, local-first agentic strategy lab in your terminal. Chat with a
+LangGraph agent to fetch keyless Binance public market data, design and backtest
+long/flat crypto strategies, and run simulated paper trades. Charts render to
+self-contained plotly HTML that opens in your browser.
+
+**Paper-only by design.** backtestchat simulates fills from closed bars. It never
+places real orders and never needs exchange API keys. Market data comes from the
+public, keyless Binance REST API. The only key you supply is your own model
+(BYOK) key, or a local Ollama endpoint.
+
+> backtestchat is a research and simulation tool and a portfolio/learning project.
+> It is not financial advice and not a trading product. Do not use it to make
+> investment decisions.
+
+<!-- TODO: asciinema GIF of a chat session (quote -> backtest -> chart). -->
+<!-- TODO: screenshot of a backtestReport chart. -->
+
+## Quickstart
+
+Requires Python 3.12+ and [uv](https://docs.astral.sh/uv/).
+
+```bash
+git clone https://github.com/barisarat/backtestchat
+cd backtestchat
+uv sync --extra openai              # deps + the OpenAI model provider
+
+export BACKTESTCHAT_MODEL=openai:gpt-5.4-nano
+export OPENAI_API_KEY=sk-...
+
+uv run backtestchat chat                # talk to the agent
+uv run backtestchat runs sync           # advance paper runs to the latest closed bar
+uv run backtestchat runs list           # check paper run status
+```
+
+Two things to know:
+
+- The `backtestchat` command lives in the project venv, so either prefix every
+  command with `uv run`, or activate the venv once with
+  `source .venv/bin/activate` and call `backtestchat` directly.
+- Always sync with the extra for the provider you use. A plain `uv sync`
+  removes extras, and `backtestchat chat` will fail until you
+  `uv sync --extra openai` again.
+
+Market data works with **zero keys**; only the chat agent needs a model key.
+Other providers (Anthropic, Ollama, ...) are covered under
+[Model setup](#model-setup-byok).
+
+## What it does
+
+- **Chat** with a tool-calling agent (LangGraph) that maps your plain-English
+  requests onto market-data, backtest, and paper-trade actions.
+- **Backtest** ~11 single-signal long/flat strategies (SMA/EMA crossover, MACD,
+  RSI, Bollinger, Donchian, ROC, stochastic, ATR channel, price-vs-SMA,
+  buy & hold) with CAGR, Sharpe, max drawdown, win rate, and trade count.
+- **Compare** several strategies at once, ranked, on one chart.
+- **Paper-trade** a strategy with simulated fills. Runs advance by replaying
+  closed bars - no daemon, no scheduler. Any command can trigger a cheap sync.
+- **Charts** are deterministic plotly HTML written to `./artifacts`. The model
+  never writes HTML; it only triggers the tools that build the charts.
+
+Supported intervals: **15m, 1h, 4h, 1d**. Symbols: any Binance spot pair
+(BTCUSDT, ETHUSDT, SOLUSDT, XRPUSDT, DOGEUSDT, ...).
+
+## Model setup (BYOK)
+
+backtestchat talks to any provider via LangChain's `init_chat_model`, configured with
+a `provider:model` string in `BACKTESTCHAT_MODEL` (or `model = "..."` in
+`./backtestchat.toml`). The default setup uses OpenAI:
+
+```bash
+uv sync --extra openai
+export BACKTESTCHAT_MODEL=openai:gpt-5.4-nano
+export OPENAI_API_KEY=sk-...
+```
+
+Any provider `init_chat_model` supports works the same way: install the matching
+extra (`--extra anthropic`, `--extra ollama`, ...) and set `BACKTESTCHAT_MODEL` to
+that provider's `provider:model` string plus its API key (Ollama runs locally
+and needs no key).
+
+If no model is configured, `backtestchat chat` prints these instructions and exits.
+Check what is resolved with `backtestchat config`.
+
+## Usage
+
+```bash
+uv run backtestchat chat
+```
+
+Then talk to it naturally:
+
+```
+you > what's BTC trading at?
+you > backtest an ma crossover 20/50 on BTCUSDT 1h over the last 3 months
+you > compare that against buy and hold and rsi
+you > save that strategy            # asks you to confirm in the terminal
+you > start a paper run with 5000 usdt   # asks you to confirm
+you > list my paper runs
+```
+
+Charts open automatically in your browser and are saved under `./artifacts`.
+Saving a strategy and starting a paper run pause for a y/n confirmation
+(human-in-the-loop, implemented with LangGraph interrupts).
+
+### Commands
+
+```bash
+backtestchat chat [--new] [--thread ID]   # agent REPL (resumes the last thread by default)
+backtestchat runs list [--status active]  # list paper runs (stored state)
+backtestchat runs sync                    # advance all active runs to the latest closed bar
+backtestchat show [last|FILE]             # open a rendered chart artifact
+backtestchat config                       # print resolved configuration
+```
+
+### How paper runs work
+
+A paper run does not backdate: the first simulated fill happens on the first bar
+that **closes after** you create the run, so a fresh run shows
+"waiting for first tick" until then. Advancing a run replays every closed bar
+since its last processed bar, one signal evaluation per bar. Because the
+already-processed-bar guard is the only cadence control, `backtestchat runs sync` is
+safe to run anytime and is idempotent.
+
+## LangSmith tracing (optional)
+
+Set the standard env vars and traces appear in your LangSmith project - no code
+or flags required:
+
+```bash
+export LANGSMITH_TRACING=true
+export LANGSMITH_API_KEY=...
+```
+
+## Configuration
+
+Resolved in order: environment variables > `./backtestchat.toml` > defaults.
+
+| Key | Env var | Default |
+|---|---|---|
+| model | `BACKTESTCHAT_MODEL` | (unset; required for chat) |
+| data dir | `BACKTESTCHAT_DATA_DIR` | `./data` |
+| artifacts dir | `BACKTESTCHAT_ARTIFACTS_DIR` | `./artifacts` |
+
+`./data` holds the sqlite store (strategies, paper runs, trades) and the
+LangGraph checkpoint database. `./artifacts` holds rendered chart HTML. Both are
+gitignored.
+
+## Development
+
+```bash
+uv sync --extra openai  # dev dependencies included; keep your provider extra
+uv run pytest -q        # deterministic, no network, no model key
+```
+
+The engine, signals, broker window logic, and paper runner are pure and tested
+against fixture bars. Market-data and agent layers are tested with mocked HTTP
+and a scripted fake model, so the whole suite runs offline.
+
+## License
+
+MIT - see [LICENSE](LICENSE).

backtestchat-0.1.0/pyproject.toml

@@ -0,0 +1,61 @@
+[project]
+name = "backtestchat"
+version = "0.1.0"
+description = "Local-first agentic strategy lab CLI: chat, backtest, and paper-trade long/flat strategies on keyless Binance public data."
+readme = "README.md"
+requires-python = ">=3.12"
+license = { text = "MIT" }
+license-files = ["LICENSE"]
+authors = [{ name = "Baris Arat" }]
+keywords = ["langgraph", "trading", "backtest", "paper-trading", "binance", "agent", "cli"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3.12",
+    "Topic :: Office/Business :: Financial :: Investment",
+    "Topic :: Scientific/Engineering",
+]
+
+dependencies = [
+    "langgraph>=0.2.60",
+    "langgraph-checkpoint-sqlite>=2.0.0",
+    "langchain>=0.3.0",
+    "langchain-core>=0.3.0",
+    "typer>=0.12.0",
+    "rich>=13.7.0",
+    "httpx>=0.27.0",
+    "plotly>=5.22.0",
+]
+
+[project.optional-dependencies]
+openai = ["langchain-openai>=0.2.0"]
+anthropic = ["langchain-anthropic>=0.2.0"]
+ollama = ["langchain-ollama>=0.2.0"]
+
+[project.scripts]
+backtestchat = "backtestchat.cli:app"
+btchat = "backtestchat.cli:app"
+
+[project.urls]
+Homepage = "https://github.com/barisarat/backtestchat"
+Repository = "https://github.com/barisarat/backtestchat"
+Issues = "https://github.com/barisarat/backtestchat/issues"
+
+[dependency-groups]
+dev = [
+    "pytest>=8.2.0",
+    "respx>=0.21.0",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/backtestchat"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+addopts = "-q"

backtestchat-0.1.0/src/backtestchat/__init__.py

@@ -0,0 +1,8 @@
+"""backtestchat: a local-first agentic strategy lab CLI.
+
+Chat with a LangGraph agent to fetch keyless Binance public market data,
+backtest long/flat strategies, and run simulated paper trades. Paper-only by
+design: no exchange keys, no real orders.
+"""
+
+__version__ = "0.1.0"

backtestchat-0.1.0/src/backtestchat/__main__.py

@@ -0,0 +1,6 @@
+"""Enable `python -m backtestchat`."""
+
+from backtestchat.cli import app
+
+if __name__ == "__main__":
+    app()