scanlang
Screener DSL and scan compiler: signal definitions to polars pushdown filters.
A scan definition is a plain dict (JSON from a UI, a Python literal from a notebook)
that scanlang compiles into one validated polars predicate. Nothing is
string-interpolated, so there is no injection surface. Filters run on eager
DataFrame or lazy LazyFrame; window semantics are computed per partition.
Status: v0.2, pre-alpha. The IR is frozen (additive changes only). For the design rationale, see docs/explanation/ir-design.md.
Install
uv add scanlang # or: pip install scanlang
Requires Python >= 3.11 and polars >= 1.44. The optional talib extra is a
placeholder for a future value-parity indicator module; nothing in the public
API depends on it yet.
Quickstart (eager)
Copy-pasteable end-to-end. Defines its own small OHLCV frame, scores it,
applies a scan, and prints the picks. The full annotated walkthrough lives in
docs/examples/01_quickstart.py.
import polars as pl
from scanlang import apply, score_bars, validate
bars = pl.DataFrame({
"symbol": ["AAA"] * 30 + ["BBB"] * 30,
"session": pl.date_range(pl.date(2026, 1, 1), pl.date(2026, 3, 1), interval="1d", eager=True)[:30].to_list() * 2,
"open": [10 + i for i in range(30)] + [60 - i for i in range(30)],
"high": [11 + i for i in range(30)] + [61 - i for i in range(30)],
"low": [9 + i for i in range(30)] + [59 - i for i in range(30)],
"close": [10 + i for i in range(30)] + [60 - i for i in range(30)],
"volume": [1000.0] * 60,
})
scored = score_bars(bars).collect() # LazyFrame -> DataFrame at the edge
scan_def = {
"filters": [{"property": "score", "op": ">=", "value": 40}],
"order_by": [{"property": "score", "dir": "desc"}],
"limit": 5,
}
validate(scan_def) # [] when valid; never raises
print(apply(scored, scan_def).select("symbol", "score", "phase"))
The full quickstart script (lazy in, collect at the edge) is
docs/examples/01_quickstart.py. Side-by-side
eager/lazy/piped/renamed modes: docs/examples/07_lazy_vs_sync.py.
Quickstart (text DSL)
Prefer a one-liner over the dict? parse turns human syntax into the same
scan dict. The golden-cross form uses cross_above so the signal only fires
when the 20-EMA actually crosses above the 50-EMA:
from scanlang import parse, validate
ir = parse("cross_above(ema(20), ema(50)) and rsi(close, 14) > 70")
validate(ir) # [] when valid
ema(20) and sma(20) imply the close column; indicators that take an
expression like rsi need it spelled out (rsi(close, 14)). Full grammar and
operator reference: docs/reference/operators.md,
docs/how-to/scan-from-text.md.
Eager vs lazy at a glance
| Mode | In | Out | When to use it |
|---|---|---|---|
| Sync / eager | pl.DataFrame |
pl.DataFrame |
Notebook, REPL, small script. Collect at the data edge, then everything stays eager. |
| Lazy end-to-end | pl.LazyFrame |
pl.LazyFrame |
Pipeline that pipes into more polars ops. Add .collect() once at the end. |
| Mixed | pl.LazyFrame (start) -> pl.DataFrame (collect once) |
pl.DataFrame |
One .collect() at the polars -> non-polars boundary. Don't collect "to be safe" earlier; you lose predicate pushdown. |
apply is shape-preserving (eager in -> eager out, lazy in -> lazy out).
score_bars is always lazy out: it returns a LazyFrame so it can fold into a
bigger polars plan; call .collect() at the edge if you want a DataFrame.
Full guide: docs/how-to/eager-frames.md.
Docs (Diataxis)
Build the site locally with uv run --group docs zensical build (config:
zensical.toml). Follows the Diataxis
split:
- Tutorials - learning-oriented; get to a first scan.
- How-to guides - task-oriented; solve a specific problem (custom catalog/partition, extending indicators, scan from text, score + stats).
- Explanation - understanding-oriented; IR design, lazy contract, null semantics, validation split, and why there is no SQL backend.
- Reference - information-oriented; API, operators, indicators, examples index, notebooks, IR freeze.
Notebooks: 01_first_scan.ipynb (Jupyter)
and 02_first_scan_marimo.py (marimo)
- the same first scan as a runnable notebook. See
docs/reference/notebooks.md.
Development
uv sync --group docs # create .venv with zensical
.venv/bin/python -m pytest tests/ -q # tests
.venv/bin/python -m ruff check src tests # lint
.venv/bin/zensical serve # live-reload docs at :8000
.venv/bin/zensical build # static build -> site/
License
MIT.
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file scanlang-0.3.0.tar.gz.
File metadata
- Download URL: scanlang-0.3.0.tar.gz
- Upload date:
- Size: 30.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
uv/0.12.9 {"installer":{"name":"uv","version":"0.12.9","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
05502e3953077451e43493256f173a2f8ce7d663b1d3e60723f369f2ee71ed65
|
|
| MD5 |
a67cfcff64db65c0ca0797ac670d1e58
|
|
| BLAKE2b-256 |
2174a8a24c6bdb3fcbfa0bd0f976deaf249d04bde0a58ab61d1068895751d312
|
File details
Details for the file scanlang-0.3.0-py3-none-any.whl.
File metadata
- Download URL: scanlang-0.3.0-py3-none-any.whl
- Upload date:
- Size: 33.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via:
uv/0.12.9 {"installer":{"name":"uv","version":"0.12.9","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
934f2781f2041d983691ed7bd4720f2158ee91a0010d64497e697431a57345cd
|
|
| MD5 |
42220a9290ccf1c2bb0b9d863fde5a7a
|
|
| BLAKE2b-256 |
fab32020fdd07cc4fdb76f86b8a92d2e87f3f8fabfb49fed1386ae9c64ef15e5
|