openstat-cli 1.0.0__py3-none-any.whl

openstat/__init__.py

@@ -0,0 +1,3 @@
+"""OpenStat — Open-source statistical analysis tool."""
+
+__version__ = "0.9.0"

openstat/__main__.py

@@ -0,0 +1,4 @@
+"""Allow running as `python -m openstat`."""
+from openstat.cli import app
+
+app()

openstat/backends/__init__.py

@@ -0,0 +1,16 @@
+"""Data backend abstraction: Polars (default) and DuckDB."""
+
+from __future__ import annotations
+
+from typing import Protocol
+
+import polars as pl
+
+
+class DataBackend(Protocol):
+    """Protocol for data backends."""
+
+    def load(self, path: str) -> None: ...
+    def to_polars(self) -> pl.DataFrame: ...
+    def sql(self, query: str) -> pl.DataFrame: ...
+    def shape(self) -> tuple[int, int]: ...

openstat/backends/duckdb_backend.py

@@ -0,0 +1,70 @@
+"""DuckDB backend for large dataset processing."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import polars as pl
+
+
+def _try_import_duckdb():
+    try:
+        import duckdb
+        return duckdb
+    except ImportError:
+        raise ImportError(
+            "DuckDB backend requires duckdb. "
+            "Install it with: pip install openstat[duckdb]"
+        )
+
+
+class DuckDBBackend:
+    """DuckDB-based backend for large datasets."""
+
+    def __init__(self) -> None:
+        duckdb = _try_import_duckdb()
+        self._conn = duckdb.connect()
+        self._table_loaded = False
+
+    def load(self, path: str) -> None:
+        p = Path(path)
+        suffix = p.suffix.lower()
+        path_str = str(p).replace("'", "''")
+
+        if suffix == ".csv":
+            self._conn.execute(f"CREATE OR REPLACE TABLE data AS SELECT * FROM read_csv('{path_str}')")
+        elif suffix == ".parquet":
+            self._conn.execute(f"CREATE OR REPLACE TABLE data AS SELECT * FROM read_parquet('{path_str}')")
+        elif suffix in (".xlsx", ".xls"):
+            # DuckDB doesn't natively support Excel; load via Polars and register
+            df = pl.read_excel(p)
+            self._conn.register("data", df.to_pandas())
+        else:
+            from openstat.io.loader import load_file
+            df = load_file(path)
+            self._conn.register("data", df.to_pandas())
+
+        self._table_loaded = True
+
+    def to_polars(self) -> pl.DataFrame:
+        if not self._table_loaded:
+            raise RuntimeError("No data loaded")
+        return self._conn.execute("SELECT * FROM data").pl()
+
+    def sql(self, query: str) -> pl.DataFrame:
+        if not self._table_loaded:
+            raise RuntimeError("No data loaded. Use 'load' first.")
+        return self._conn.execute(query).pl()
+
+    def shape(self) -> tuple[int, int]:
+        if not self._table_loaded:
+            return (0, 0)
+        result = self._conn.execute("SELECT COUNT(*) as n FROM data").fetchone()
+        n_rows = result[0] if result else 0
+        cols = self._conn.execute("SELECT * FROM data LIMIT 0").description
+        n_cols = len(cols) if cols else 0
+        return (n_rows, n_cols)
+
+    def execute(self, query: str):
+        """Execute raw SQL and return DuckDB result."""
+        return self._conn.execute(query)

openstat/backends/polars_backend.py

@@ -0,0 +1,52 @@
+"""Polars backend — wraps the default Polars DataFrame operations."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import polars as pl
+
+from openstat.config import get_config
+
+
+class PolarsBackend:
+    """Default backend using Polars DataFrames."""
+
+    def __init__(self) -> None:
+        self._df: pl.DataFrame | None = None
+        self._lazy: pl.LazyFrame | None = None
+
+    def load(self, path: str) -> None:
+        p = Path(path)
+        suffix = p.suffix.lower()
+        cfg = get_config()
+
+        # Use LazyFrame for scan-capable formats
+        if suffix == ".csv":
+            self._lazy = pl.scan_csv(
+                p,
+                infer_schema_length=cfg.infer_schema_length,
+            )
+        elif suffix == ".parquet":
+            self._lazy = pl.scan_parquet(p)
+        else:
+            # For other formats, load eagerly
+            from openstat.io.loader import load_file
+            self._df = load_file(path)
+            self._lazy = self._df.lazy()
+
+    def to_polars(self) -> pl.DataFrame:
+        if self._df is not None:
+            return self._df
+        if self._lazy is not None:
+            self._df = self._lazy.collect()
+            return self._df
+        raise RuntimeError("No data loaded")
+
+    def sql(self, query: str) -> pl.DataFrame:
+        df = self.to_polars()
+        return pl.SQLContext({"data": df}).execute(query).collect()
+
+    def shape(self) -> tuple[int, int]:
+        df = self.to_polars()
+        return df.shape

openstat/cli.py

@@ -0,0 +1,92 @@
+"""CLI entry point for OpenStat (Typer-based)."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import typer
+from rich.console import Console
+
+from openstat import __version__
+
+app = typer.Typer(
+    name="openstat",
+    help="OpenStat — Open-source statistical analysis tool.",
+    add_completion=False,
+    no_args_is_help=True,
+)
+console = Console()
+
+
+def _version_callback(value: bool) -> None:
+    if value:
+        console.print(f"openstat {__version__}")
+        raise typer.Exit()
+
+
+@app.callback()
+def main(
+    version: bool = typer.Option(
+        False, "--version", "-V",
+        callback=_version_callback,
+        is_eager=True,
+        help="Show version and exit.",
+    ),
+    verbose: bool = typer.Option(
+        False, "--verbose", "-v",
+        help="Enable verbose logging (INFO level).",
+    ),
+    debug: bool = typer.Option(
+        False, "--debug",
+        help="Enable debug logging (DEBUG level).",
+    ),
+) -> None:
+    """OpenStat — Open-source statistical analysis tool."""
+    from openstat.logging_config import setup_logging
+    setup_logging(verbose=verbose, debug=debug)
+
+
+@app.command()
+def repl() -> None:
+    """Start the interactive REPL."""
+    from openstat.repl import run_repl
+    run_repl()
+
+
+@app.command()
+def run(
+    script: Path = typer.Argument(..., help="Path to an .ost script file."),
+    strict: bool = typer.Option(
+        False, "--strict",
+        help="Stop on first error and exit with code 1.",
+    ),
+) -> None:
+    """Execute an .ost script file."""
+    if not script.exists():
+        console.print(f"[red]File not found: {script}[/red]")
+        raise typer.Exit(1)
+    from openstat.repl import run_script
+    run_script(str(script), strict=strict)
+
+
+@app.command()
+def serve(
+    port: int = typer.Option(8080, help="Port to listen on."),
+    host: str = typer.Option("127.0.0.1", help="Host to bind to."),
+) -> None:
+    """Start the web-based GUI."""
+    try:
+        import uvicorn
+    except ImportError:
+        console.print("[red]Web GUI requires: pip install openstat[web][/red]")
+        raise typer.Exit(1)
+    from openstat.web.app import app as web_app
+    if web_app is None:
+        console.print("[red]Web GUI requires: pip install openstat[web][/red]")
+        raise typer.Exit(1)
+    console.print(f"[bold cyan]OpenStat Web[/bold cyan] starting on http://{host}:{port}")
+    uvicorn.run(web_app, host=host, port=port, log_level="info")
+
+
+if __name__ == "__main__":
+    app()

openstat/commands/__init__.py

@@ -0,0 +1,82 @@
+"""Command package — import all command modules to register them."""
+
+# Importing these modules triggers the @command decorators,
+# which populate the global registry in base.py.
+from openstat.commands import data_cmds   # noqa: F401
+from openstat.commands import stat_cmds   # noqa: F401
+from openstat.commands import plot_cmds   # noqa: F401
+from openstat.commands import report_cmds  # noqa: F401
+from openstat.commands import plugin_cmds  # noqa: F401
+from openstat.commands import panel_cmds   # noqa: F401
+from openstat.commands import iv_cmds      # noqa: F401
+from openstat.commands import mixed_cmds   # noqa: F401
+from openstat.commands import ts_cmds      # noqa: F401
+from openstat.commands import surv_cmds    # noqa: F401
+from openstat.commands import mi_cmds      # noqa: F401
+from openstat.commands import survey_cmds  # noqa: F401
+from openstat.commands import backend_cmds  # noqa: F401
+from openstat.commands import discrete_cmds  # noqa: F401
+from openstat.commands import causal_cmds    # noqa: F401
+from openstat.commands import power_cmds     # noqa: F401
+from openstat.commands import factor_cmds    # noqa: F401
+from openstat.commands import export_cmds    # noqa: F401
+from openstat.commands import nonparam_cmds  # noqa: F401
+from openstat.commands import ml_cmds        # noqa: F401
+from openstat.commands import cluster_cmds   # noqa: F401
+from openstat.commands import manova_cmds    # noqa: F401
+from openstat.commands import arch_cmds      # noqa: F401
+from openstat.commands import bayes_cmds     # noqa: F401
+from openstat.commands import reshape_cmds   # noqa: F401
+from openstat.commands import advreg_cmds    # noqa: F401
+from openstat.commands import ts_adv_cmds   # noqa: F401
+from openstat.commands import influence_cmds # noqa: F401
+from openstat.commands import ml_adv_cmds   # noqa: F401
+from openstat.commands import esttab_cmds   # noqa: F401
+from openstat.commands import string_cmds   # noqa: F401
+from openstat.commands import epi_cmds      # noqa: F401
+from openstat.commands import dsl_cmds          # noqa: F401
+from openstat.commands import resampling_cmds   # noqa: F401
+from openstat.commands import model_eval_cmds   # noqa: F401
+from openstat.commands import dataquality_cmds  # noqa: F401
+from openstat.commands import outreg_cmds       # noqa: F401
+from openstat.commands import equiv_tobit_cmds  # noqa: F401
+from openstat.commands import viz_extra_cmds    # noqa: F401
+from openstat.commands import posthoc_cmds      # noqa: F401
+from openstat.commands import sem_cmds          # noqa: F401
+from openstat.commands import meta_cmds         # noqa: F401
+from openstat.commands import network_cmds      # noqa: F401
+from openstat.commands import automodel_cmds    # noqa: F401
+from openstat.commands import repro_cmds        # noqa: F401
+from openstat.commands import tui_cmds          # noqa: F401
+from openstat.commands import i18n_cmds         # noqa: F401
+from openstat.commands import mediate_cmds      # noqa: F401
+from openstat.commands import validate_cmds     # noqa: F401
+from openstat.commands import regex_cmds        # noqa: F401
+from openstat.commands import alias_cmds        # noqa: F401
+from openstat.commands import pdf_cmds          # noqa: F401
+from openstat.commands import watch_cmds        # noqa: F401
+from openstat.commands import rbridge_cmds      # noqa: F401
+from openstat.commands import nlquery_cmds      # noqa: F401
+from openstat.commands import stata_import_cmds # noqa: F401
+from openstat.commands import help_cmds         # noqa: F401
+from openstat.commands import datetime_cmds     # noqa: F401
+from openstat.commands import groupby_cmds      # noqa: F401
+from openstat.commands import profile_cmds      # noqa: F401
+from openstat.commands import advanced_ml_cmds  # noqa: F401
+from openstat.commands import viz_adv_cmds      # noqa: F401
+from openstat.commands import pipeline_cmds     # noqa: F401
+from openstat.commands import export_extra_cmds # noqa: F401
+from openstat.commands import datamanip_cmds    # noqa: F401
+from openstat.commands import import_extra_cmds # noqa: F401
+from openstat.commands import adv_stat_cmds      # noqa: F401
+from openstat.commands import mixture_changepoint_cmds  # noqa: F401
+from openstat.commands import dimreduce_cmds     # noqa: F401
+from openstat.commands import arules_cmds        # noqa: F401
+from openstat.commands import textanalysis_cmds  # noqa: F401
+from openstat.commands import ux_cmds            # noqa: F401
+from openstat.commands import export_beamer_cmds # noqa: F401
+
+from openstat.commands.base import get_registry
+
+# Public API — the COMMANDS dict used by the REPL
+COMMANDS = get_registry()