pipeverify 0.1.0__py3-none-any.whl

pipeverify/__init__.py

@@ -0,0 +1,8 @@
+"""pipeverify: check that a pipeline change left the data right, not just the job green."""
+
+__version__ = "0.1.0"
+
+from .checks import Finding, Thresholds  # noqa: E402
+from .core import RunResult, TableResult, run_config, verify  # noqa: E402
+
+__all__ = ["verify", "run_config", "Thresholds", "Finding", "TableResult", "RunResult", "__version__"]

pipeverify/checks.py

@@ -0,0 +1,285 @@
+"""Checks that compare a table before and after a change.
+
+Each check looks for a kind of silent harm: the job succeeds and tests stay green,
+but the data is no longer right. Checks return Findings; the policy turns them
+into a verdict.
+"""
+from __future__ import annotations
+
+from dataclasses import asdict, dataclass, field
+from typing import Any
+
+import duckdb
+
+PASS, INFO, REVIEW, BLOCK = "pass", "info", "review", "block"
+RANK = {PASS: 0, INFO: 0, REVIEW: 1, BLOCK: 2}
+
+NUMERIC_TYPES = ("TINYINT", "SMALLINT", "INTEGER", "BIGINT", "HUGEINT", "UTINYINT", "USMALLINT",
+                 "UINTEGER", "UBIGINT", "FLOAT", "REAL", "DOUBLE", "DECIMAL")
+TEXT_TYPES = ("VARCHAR", "TEXT", "STRING", "CHAR", "ENUM")
+
+
+@dataclass
+class Thresholds:
+    """When a difference becomes worth a human look (review) or should stop the change (block).
+
+    Percentages are shares of rows, from 0 to 100.
+    """
+
+    row_change_review_pct: float = 1.0
+    row_change_block_pct: float = 10.0
+    missing_keys_review_pct: float = 0.0     # any lost key gets a review
+    missing_keys_block_pct: float = 5.0
+    null_rate_review_pts: float = 1.0
+    null_rate_block_pts: float = 10.0
+    changed_values_review_pct: float = 0.0   # any unexpected changed value gets a review
+    changed_values_block_pct: float = 10.0
+    sum_shift_review_pct: float = 0.5
+    sum_shift_block_pct: float = 5.0
+    category_max_distinct: int = 50          # treat text columns with few values as categories
+    float_tolerance: float = 1e-9
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any] | None) -> "Thresholds":
+        if not data:
+            return cls()
+        known = {f for f in cls.__dataclass_fields__}
+        unknown = set(data) - known
+        if unknown:
+            raise ValueError(f"Unknown threshold(s): {', '.join(sorted(unknown))}")
+        return cls(**data)
+
+
+@dataclass
+class Finding:
+    check: str
+    severity: str
+    message: str
+    column: str | None = None
+    metrics: dict[str, Any] = field(default_factory=dict)
+
+    def to_dict(self) -> dict[str, Any]:
+        return asdict(self)
+
+
+def q(name: str) -> str:
+    return '"' + name.replace('"', '""') + '"'
+
+
+def _grade(value: float, review: float, block: float) -> str:
+    if value >= block and value > 0:
+        return BLOCK
+    if value > review or (review == 0 and value > 0):
+        return REVIEW
+    return PASS
+
+
+def _schema(con: duckdb.DuckDBPyConnection, table: str) -> dict[str, str]:
+    rows = con.execute(f"DESCRIBE {q(table)}").fetchall()
+    return {r[0]: str(r[1]).upper() for r in rows}
+
+
+def _is_numeric(dtype: str) -> bool:
+    return dtype.startswith(NUMERIC_TYPES)
+
+
+def _is_text(dtype: str) -> bool:
+    return dtype.startswith(TEXT_TYPES)
+
+
+def run_checks(
+    con: duckdb.DuckDBPyConnection,
+    before: str,
+    after: str,
+    key: list[str] | None = None,
+    expect_change: list[str] | None = None,
+    ignore: list[str] | None = None,
+    expect_row_change: bool = False,
+    thresholds: Thresholds | None = None,
+) -> tuple[list[Finding], dict[str, Any]]:
+    """Run every check on two tables already loaded in `con`."""
+    t = thresholds or Thresholds()
+    key = list(key or [])
+    expect = set(expect_change or [])
+    ignore = set(ignore or [])
+    findings: list[Finding] = []
+
+    sb, sa = _schema(con, before), _schema(con, after)
+    for k in key:
+        if k not in sb or k not in sa:
+            raise ValueError(f"Key column '{k}' is missing from the before or after table")
+
+    n_before = con.execute(f"SELECT count(*) FROM {q(before)}").fetchone()[0]
+    n_after = con.execute(f"SELECT count(*) FROM {q(after)}").fetchone()[0]
+    stats: dict[str, Any] = {"rows_before": n_before, "rows_after": n_after,
+                             "columns_before": len(sb), "columns_after": len(sa)}
+
+    # 1. schema
+    removed = [c for c in sb if c not in sa and c not in ignore]
+    added = [c for c in sa if c not in sb and c not in ignore]
+    retyped = [c for c in sb if c in sa and c not in ignore and sb[c] != sa[c]]
+    for c in removed:
+        findings.append(Finding("schema", BLOCK, f"Column '{c}' was removed", c, {"type": sb[c]}))
+    for c in added:
+        findings.append(Finding("schema", INFO, f"Column '{c}' was added", c, {"type": sa[c]}))
+    for c in retyped:
+        findings.append(Finding("schema", REVIEW, f"Column '{c}' changed type {sb[c]} -> {sa[c]}",
+                                c, {"before": sb[c], "after": sa[c]}))
+
+    # 2. row count
+    if n_before == 0 and n_after == 0:
+        findings.append(Finding("row_count", INFO, "Both tables are empty"))
+    else:
+        base = max(n_before, 1)
+        pct = abs(n_after - n_before) / base * 100
+        sev = INFO if expect_row_change else _grade(pct, t.row_change_review_pct, t.row_change_block_pct)
+        if n_before > 0 and n_after == 0:
+            sev = BLOCK
+        if n_after != n_before:
+            direction = "lost" if n_after < n_before else "gained"
+            findings.append(Finding("row_count", sev,
+                                    f"Table {direction} {abs(n_after - n_before):,} rows ({pct:.2f}%)",
+                                    None, {"before": n_before, "after": n_after, "pct": round(pct, 4)}))
+
+    common = [c for c in sb if c in sa and c not in ignore]
+
+    # 3. keys
+    unique_keys_join = None
+    if key:
+        kcols = ", ".join(q(k) for k in key)
+        dup_before = con.execute(
+            f"SELECT count(*) FROM (SELECT {kcols} FROM {q(before)} GROUP BY ALL HAVING count(*) > 1)").fetchone()[0]
+        dup_after = con.execute(
+            f"SELECT count(*) FROM (SELECT {kcols} FROM {q(after)} GROUP BY ALL HAVING count(*) > 1)").fetchone()[0]
+        stats.update(duplicate_keys_before=dup_before, duplicate_keys_after=dup_after)
+        if dup_after > dup_before:
+            findings.append(Finding("duplicate_keys", BLOCK,
+                                    f"{dup_after - dup_before:,} key value(s) now appear more than once",
+                                    None, {"before": dup_before, "after": dup_after}))
+        elif dup_after < dup_before:
+            findings.append(Finding("duplicate_keys", INFO,
+                                    f"Duplicate keys went down from {dup_before:,} to {dup_after:,}",
+                                    None, {"before": dup_before, "after": dup_after}))
+        on = " AND ".join(f"b.{q(k)} IS NOT DISTINCT FROM a.{q(k)}" for k in key)
+        missing = con.execute(
+            f"SELECT count(*) FROM (SELECT DISTINCT {kcols} FROM {q(before)}) b "
+            f"WHERE NOT EXISTS (SELECT 1 FROM {q(after)} a WHERE {on})").fetchone()[0]
+        new = con.execute(
+            f"SELECT count(*) FROM (SELECT DISTINCT {kcols} FROM {q(after)}) a "
+            f"WHERE NOT EXISTS (SELECT 1 FROM {q(before)} b WHERE {on})").fetchone()[0]
+        distinct_before = con.execute(f"SELECT count(*) FROM (SELECT DISTINCT {kcols} FROM {q(before)})").fetchone()[0]
+        stats.update(missing_keys=missing, new_keys=new)
+        if missing:
+            pct = missing / max(distinct_before, 1) * 100
+            sev = INFO if expect_row_change else _grade(pct, t.missing_keys_review_pct, t.missing_keys_block_pct)
+            findings.append(Finding("missing_keys", sev,
+                                    f"{missing:,} key(s) present before are gone after ({pct:.2f}%)",
+                                    None, {"count": missing, "pct": round(pct, 4)}))
+        if new:
+            findings.append(Finding("new_keys", INFO if expect_row_change else REVIEW,
+                                    f"{new:,} new key(s) appeared", None, {"count": new}))
+        # rows whose key is unique on both sides, used for value comparison
+        con.execute(f"CREATE OR REPLACE TEMP VIEW _pv_ub AS SELECT * FROM {q(before)} "
+                    f"QUALIFY count(*) OVER (PARTITION BY {kcols}) = 1")
+        con.execute(f"CREATE OR REPLACE TEMP VIEW _pv_ua AS SELECT * FROM {q(after)} "
+                    f"QUALIFY count(*) OVER (PARTITION BY {kcols}) = 1")
+        unique_keys_join = " AND ".join(f"b.{q(k)} IS NOT DISTINCT FROM a.{q(k)}" for k in key)
+
+    # 4. null rates
+    for c in common:
+        nb, na = con.execute(
+            f"SELECT (SELECT avg(CASE WHEN {q(c)} IS NULL THEN 1.0 ELSE 0 END) FROM {q(before)}), "
+            f"(SELECT avg(CASE WHEN {q(c)} IS NULL THEN 1.0 ELSE 0 END) FROM {q(after)})").fetchone()
+        nb, na = float(nb or 0) * 100, float(na or 0) * 100
+        rise = na - nb
+        if rise > 0:
+            sev = _grade(rise, t.null_rate_review_pts, t.null_rate_block_pts)
+            if c in expect and sev != PASS:
+                sev = INFO
+            if sev != PASS:
+                findings.append(Finding("null_rate", sev,
+                                        f"Nulls in '{c}' rose from {nb:.2f}% to {na:.2f}%",
+                                        c, {"before_pct": round(nb, 4), "after_pct": round(na, 4)}))
+
+    # 5. changed values on matched keys
+    if unique_keys_join is not None:
+        matched = con.execute(f"SELECT count(*) FROM _pv_ub b JOIN _pv_ua a ON {unique_keys_join}").fetchone()[0]
+        stats["matched_rows"] = matched
+        for c in common:
+            if c in key:
+                continue
+            if _is_numeric(sb[c]) and _is_numeric(sa[c]):
+                diff = (f"(b.{q(c)} IS NULL) <> (a.{q(c)} IS NULL) OR "
+                        f"abs(CAST(a.{q(c)} AS DOUBLE) - CAST(b.{q(c)} AS DOUBLE)) > "
+                        f"{t.float_tolerance} * greatest(1.0, abs(CAST(b.{q(c)} AS DOUBLE)))")
+            else:
+                diff = f"CAST(b.{q(c)} AS VARCHAR) IS DISTINCT FROM CAST(a.{q(c)} AS VARCHAR)"
+            changed = con.execute(
+                f"SELECT count(*) FROM _pv_ub b JOIN _pv_ua a ON {unique_keys_join} WHERE {diff}").fetchone()[0]
+            if not changed:
+                continue
+            pct = changed / max(matched, 1) * 100
+            if c in expect:
+                sev = INFO
+            else:
+                sev = _grade(pct, t.changed_values_review_pct, t.changed_values_block_pct)
+            sample = con.execute(
+                f"SELECT {', '.join('b.' + q(k) for k in key)}, b.{q(c)} AS before_value, a.{q(c)} AS after_value "
+                f"FROM _pv_ub b JOIN _pv_ua a ON {unique_keys_join} WHERE {diff} LIMIT 3").fetchall()
+            findings.append(Finding(
+                "changed_values", sev,
+                f"{changed:,} row(s) changed '{c}' ({pct:.2f}% of matched rows)"
+                + ("" if c not in expect else ", as expected"),
+                c, {"count": changed, "pct": round(pct, 4),
+                    "examples": [[str(v) for v in row] for row in sample]}))
+
+    # 6. totals of numeric columns
+    for c in common:
+        if c in key or not (_is_numeric(sb[c]) and _is_numeric(sa[c])):
+            continue
+        tb, ta = con.execute(
+            f"SELECT (SELECT sum(CAST({q(c)} AS DOUBLE)) FROM {q(before)}), "
+            f"(SELECT sum(CAST({q(c)} AS DOUBLE)) FROM {q(after)})").fetchone()
+        if tb is None and ta is None:
+            continue
+        tb, ta = float(tb or 0), float(ta or 0)
+        if tb == ta:
+            continue
+        pct = abs(ta - tb) / max(abs(tb), 1e-12) * 100
+        if c in expect or expect_row_change:
+            sev = INFO
+        else:
+            sev = _grade(pct, t.sum_shift_review_pct, t.sum_shift_block_pct)
+        if sev != PASS:
+            findings.append(Finding("total_shift", sev,
+                                    f"Total of '{c}' moved {pct:.2f}% ({tb:,.2f} -> {ta:,.2f})",
+                                    c, {"before": tb, "after": ta, "pct": round(pct, 4)}))
+
+    # 7. category values
+    for c in common:
+        if c in key or not (_is_text(sb[c]) and _is_text(sa[c])):
+            continue
+        d = con.execute(f"SELECT count(DISTINCT {q(c)}) FROM {q(before)}").fetchone()[0]
+        if d == 0 or d > t.category_max_distinct:
+            continue
+        new_vals = [r[0] for r in con.execute(
+            f"SELECT DISTINCT {q(c)} FROM {q(after)} WHERE {q(c)} IS NOT NULL EXCEPT "
+            f"SELECT DISTINCT {q(c)} FROM {q(before)} ORDER BY 1 LIMIT 10").fetchall()]
+        gone_vals = [r[0] for r in con.execute(
+            f"SELECT DISTINCT {q(c)} FROM {q(before)} WHERE {q(c)} IS NOT NULL EXCEPT "
+            f"SELECT DISTINCT {q(c)} FROM {q(after)} ORDER BY 1 LIMIT 10").fetchall()]
+        if new_vals:
+            findings.append(Finding("category_values", INFO if c in expect else REVIEW,
+                                    f"'{c}' has new value(s): {', '.join(map(str, new_vals))}",
+                                    c, {"new": new_vals}))
+        if gone_vals:
+            findings.append(Finding("category_values", INFO,
+                                    f"'{c}' no longer has value(s): {', '.join(map(str, gone_vals))}",
+                                    c, {"gone": gone_vals}))
+
+    return findings, stats
+
+
+def verdict(findings: list[Finding]) -> str:
+    worst = max((RANK[f.severity] for f in findings), default=0)
+    return {0: PASS, 1: REVIEW, 2: BLOCK}[worst]

pipeverify/cli.py

@@ -0,0 +1,159 @@
+"""Command line interface: pipeverify compare | run | init."""
+from __future__ import annotations
+
+import os
+import sys
+
+import click
+
+from . import __version__
+from .checks import RANK
+from .core import RunResult, run_config, verify
+from .report import to_json, to_markdown, to_text
+
+SAMPLE_CONFIG = """\
+# pipeverify config: one entry per table you want to check after a change.
+# before/after can be a file (csv, parquet, json), duckdb:<file.db>:<table>,
+# or a mapping such as {url: postgresql://..., table: schema.table}.
+thresholds:
+  row_change_block_pct: 10
+tables:
+  - name: gold.daily_sales
+    before: snapshots/before/daily_sales.parquet
+    after: snapshots/after/daily_sales.parquet
+    key: [store_id, sale_date]
+    expect_change: []        # columns this change is meant to modify
+    ignore: [loaded_at]      # columns that always change, such as load timestamps
+"""
+
+
+def _emit(result: RunResult, fmt: str, output: str | None, summary: str | None, show_info: bool,
+          json_out: str | None = None, markdown_out: str | None = None) -> None:
+    for path, render in ((json_out, to_json), (markdown_out, to_markdown)):
+        if path:
+            with open(path, "w", encoding="utf-8") as fh:
+                body = render(result)
+                fh.write(body if body.endswith("\n") else body + "\n")
+    text = {"text": lambda: to_text(result, show_info), "markdown": lambda: to_markdown(result),
+            "json": lambda: to_json(result)}[fmt]()
+    if output:
+        with open(output, "w", encoding="utf-8") as fh:
+            fh.write(text if text.endswith("\n") else text + "\n")
+        if fmt != "text":
+            click.echo(to_text(result, show_info))
+    else:
+        click.echo(text)
+    if summary:
+        with open(summary, "a", encoding="utf-8") as fh:
+            fh.write(to_markdown(result))
+
+
+def _maybe_explain(result: RunResult, explain: bool) -> None:
+    if not explain:
+        return
+    from .explain import ExplainError, explain as do_explain
+    try:
+        result.explanation = do_explain(result)
+    except ExplainError as exc:
+        click.echo(f"note: explanation skipped: {exc}", err=True)
+    except Exception as exc:  # network or provider errors must not hide the verdict
+        click.echo(f"note: explanation failed: {exc}", err=True)
+
+
+def _exit(result: RunResult, fail_on: str) -> None:
+    if fail_on == "never":
+        sys.exit(0)
+    threshold = RANK[fail_on]
+    sys.exit(1 if RANK[result.verdict] >= threshold else 0)
+
+
+output_options = [
+    click.option("--format", "fmt", type=click.Choice(["text", "markdown", "json"]), default="text",
+                 show_default=True, help="Report format."),
+    click.option("--output", "-o", type=click.Path(dir_okay=False), help="Write the report to a file."),
+    click.option("--summary", type=click.Path(dir_okay=False), envvar="GITHUB_STEP_SUMMARY",
+                 help="Append a Markdown report here (set automatically in GitHub Actions)."),
+    click.option("--fail-on", type=click.Choice(["review", "block", "never"]), default="block",
+                 show_default=True, help="Exit with code 1 at this verdict or worse."),
+    click.option("--explain", is_flag=True, help="Add a plain-English explanation from an LLM (needs an API key)."),
+    click.option("--show-info", is_flag=True, help="Also print informational findings."),
+    click.option("--json-out", type=click.Path(dir_okay=False), help="Also write a JSON report here."),
+    click.option("--markdown-out", type=click.Path(dir_okay=False), help="Also write a Markdown report here."),
+]
+
+
+def with_output_options(fn):
+    for opt in reversed(output_options):
+        fn = opt(fn)
+    return fn
+
+
+@click.group()
+@click.version_option(__version__, prog_name="pipeverify")
+def main() -> None:
+    """Verify that a pipeline change left the data right, not just the job green."""
+
+
+@main.command()
+@click.argument("before")
+@click.argument("after")
+@click.option("--key", "-k", help="Comma-separated key column(s) that identify a row.")
+@click.option("--expect-change", "-e", multiple=True, help="Column this change is meant to modify. Repeatable.")
+@click.option("--ignore", "-i", multiple=True, help="Column to skip, such as a load timestamp. Repeatable.")
+@click.option("--expect-row-change", is_flag=True, help="Rows are expected to be added or removed.")
+@click.option("--name", default=None, help="Name to show for the table.")
+@with_output_options
+def compare(before, after, key, expect_change, ignore, expect_row_change, name,
+            fmt, output, summary, fail_on, explain, show_info, json_out, markdown_out):
+    """Compare one table BEFORE and AFTER a change.
+
+    \b
+    Examples:
+      pipeverify compare before.parquet after.parquet --key order_id
+      pipeverify compare duckdb:prod.db:gold.sales duckdb:dev.db:gold.sales -k id
+    """
+    import time
+    start = time.perf_counter()
+    try:
+        res = verify(before, after, name=name or os.path.basename(str(after)), key=key,
+                     expect_change=list(expect_change), ignore=list(ignore),
+                     expect_row_change=expect_row_change)
+    except Exception as exc:
+        raise click.ClickException(str(exc))
+    result = RunResult(tables=[res], seconds=time.perf_counter() - start)
+    _maybe_explain(result, explain)
+    _emit(result, fmt, output, summary, show_info, json_out, markdown_out)
+    _exit(result, fail_on)
+
+
+@main.command()
+@click.option("--config", "-c", default="pipeverify.yml", show_default=True,
+              type=click.Path(exists=True, dir_okay=False), help="Config file listing the tables.")
+@click.option("--table", "-t", multiple=True, help="Only check these table names. Repeatable.")
+@with_output_options
+def run(config, table, fmt, output, summary, fail_on, explain, show_info, json_out, markdown_out):
+    """Check every table listed in a config file."""
+    try:
+        result = run_config(config, only=list(table) or None)
+    except Exception as exc:
+        raise click.ClickException(str(exc))
+    if not result.tables:
+        raise click.ClickException("No tables matched")
+    _maybe_explain(result, explain)
+    _emit(result, fmt, output, summary, show_info, json_out, markdown_out)
+    _exit(result, fail_on)
+
+
+@main.command()
+@click.option("--path", default="pipeverify.yml", show_default=True)
+def init(path):
+    """Write a starter config file."""
+    if os.path.exists(path):
+        raise click.ClickException(f"{path} already exists")
+    with open(path, "w", encoding="utf-8") as fh:
+        fh.write(SAMPLE_CONFIG)
+    click.echo(f"Wrote {path}. Edit the table entries, then run: pipeverify run")
+
+
+if __name__ == "__main__":  # pragma: no cover
+    main()

pipeverify/core.py

@@ -0,0 +1,112 @@
+"""Public API: verify one table or a whole config of tables."""
+from __future__ import annotations
+
+import os
+import time
+from dataclasses import dataclass, field
+from typing import Any
+
+import duckdb
+import yaml
+
+from .checks import BLOCK, PASS, RANK, REVIEW, Finding, Thresholds, run_checks, verdict
+from .sources import load, parse_source
+
+
+@dataclass
+class TableResult:
+    name: str
+    verdict: str
+    findings: list[Finding]
+    stats: dict[str, Any]
+    before: str
+    after: str
+    error: str | None = None
+
+    def to_dict(self) -> dict[str, Any]:
+        return {"name": self.name, "verdict": self.verdict, "before": self.before, "after": self.after,
+                "stats": self.stats, "error": self.error,
+                "findings": [f.to_dict() for f in self.findings]}
+
+
+@dataclass
+class RunResult:
+    tables: list[TableResult] = field(default_factory=list)
+    seconds: float = 0.0
+    explanation: str | None = None
+
+    @property
+    def verdict(self) -> str:
+        worst = max((RANK[t.verdict] for t in self.tables), default=0)
+        return {0: PASS, 1: REVIEW, 2: BLOCK}[worst]
+
+    def to_dict(self) -> dict[str, Any]:
+        from . import __version__
+        return {"tool": "pipeverify", "version": __version__, "verdict": self.verdict,
+                "seconds": round(self.seconds, 3), "explanation": self.explanation,
+                "tables": [t.to_dict() for t in self.tables]}
+
+
+def verify(before: Any, after: Any, *, name: str = "table", key: list[str] | str | None = None,
+           expect_change: list[str] | None = None, ignore: list[str] | None = None,
+           expect_row_change: bool = False, thresholds: Thresholds | dict | None = None,
+           base_dir: str = ".") -> TableResult:
+    """Compare two snapshots of one table.
+
+    `before` and `after` can be file paths, duckdb:<file>:<table> strings,
+    config mappings, or pandas DataFrames.
+    """
+    if isinstance(key, str):
+        key = [k.strip() for k in key.split(",") if k.strip()]
+    if not isinstance(thresholds, Thresholds):
+        thresholds = Thresholds.from_dict(thresholds)
+    src_b, src_a = parse_source(before, base_dir), parse_source(after, base_dir)
+    con = duckdb.connect()
+    try:
+        load(con, src_b, "before")
+        load(con, src_a, "after")
+        findings, stats = run_checks(con, "before", "after", key=key, expect_change=expect_change,
+                                     ignore=ignore, expect_row_change=expect_row_change,
+                                     thresholds=thresholds)
+        if not key:
+            findings.append(Finding("keys", "info",
+                                    "No key given, so row-level changes were not compared. Add a key for stronger checks."))
+    finally:
+        con.close()
+    order = {BLOCK: 0, REVIEW: 1, "info": 2, PASS: 3}
+    findings.sort(key=lambda f: order[f.severity])
+    return TableResult(name=name, verdict=verdict(findings), findings=findings, stats=stats,
+                       before=src_b.describe(), after=src_a.describe())
+
+
+def load_config(path: str) -> dict[str, Any]:
+    with open(path, encoding="utf-8") as fh:
+        cfg = yaml.safe_load(fh) or {}
+    if not isinstance(cfg, dict) or "tables" not in cfg:
+        raise ValueError(f"{path}: config needs a 'tables' list")
+    return cfg
+
+
+def run_config(path: str, only: list[str] | None = None) -> RunResult:
+    cfg = load_config(path)
+    base_dir = os.path.dirname(os.path.abspath(path))
+    defaults = cfg.get("thresholds") or {}
+    result = RunResult()
+    start = time.perf_counter()
+    for spec in cfg["tables"]:
+        name = spec.get("name") or spec.get("after")
+        if only and name not in only:
+            continue
+        th = {**defaults, **(spec.get("thresholds") or {})}
+        try:
+            res = verify(spec["before"], spec["after"], name=str(name), key=spec.get("key"),
+                         expect_change=spec.get("expect_change"), ignore=spec.get("ignore"),
+                         expect_row_change=bool(spec.get("expect_row_change", False)),
+                         thresholds=th, base_dir=base_dir)
+        except Exception as exc:  # report the table as blocked, keep checking the rest
+            res = TableResult(name=str(name), verdict=BLOCK, stats={},
+                              findings=[Finding("error", BLOCK, f"Could not check this table: {exc}")],
+                              before=str(spec.get("before")), after=str(spec.get("after")), error=str(exc))
+        result.tables.append(res)
+    result.seconds = time.perf_counter() - start
+    return result

pipeverify/explain.py

@@ -0,0 +1,70 @@
+"""Optional plain-English explanation of a result by an LLM. Off unless asked for.
+
+Only the findings are sent: check names, counts, percentages and column names.
+Example values from the data are removed first unless `include_examples` is set.
+"""
+from __future__ import annotations
+
+import json
+import os
+
+from .core import RunResult
+
+PROMPT = """You review data changes for a data engineering team.
+Below is the output of pipeverify, which compared tables before and after a pipeline change.
+Write a short explanation (at most 120 words) for the engineer who made the change:
+1. What most likely happened to the data, in plain words.
+2. Which finding matters most and why.
+3. One concrete thing to check or fix next.
+Do not invent numbers. If the verdict is PASS, say so in one sentence.
+
+Result:
+{payload}
+"""
+
+
+class ExplainError(RuntimeError):
+    pass
+
+
+def _payload(result: RunResult, include_examples: bool) -> str:
+    data = result.to_dict()
+    data.pop("explanation", None)
+    for t in data["tables"]:
+        for f in t["findings"]:
+            if not include_examples:
+                f.get("metrics", {}).pop("examples", None)
+    return json.dumps(data, default=str)
+
+
+def explain(result: RunResult, provider: str | None = None, model: str | None = None,
+            include_examples: bool = False) -> str:
+    provider = (provider or os.environ.get("PIPEVERIFY_LLM_PROVIDER") or "").lower()
+    if not provider:
+        provider = "anthropic" if os.environ.get("ANTHROPIC_API_KEY") else (
+            "openai" if os.environ.get("OPENAI_API_KEY") else "")
+    if not provider:
+        raise ExplainError("Set ANTHROPIC_API_KEY or OPENAI_API_KEY to use --explain")
+    prompt = PROMPT.format(payload=_payload(result, include_examples))
+    model = model or os.environ.get("PIPEVERIFY_LLM_MODEL")
+    if provider == "anthropic":
+        try:
+            import anthropic
+        except ImportError as exc:
+            raise ExplainError("Install the AI extra: pip install 'pipeverify[ai]'") from exc
+        client = anthropic.Anthropic()
+        msg = client.messages.create(model=model or "claude-sonnet-5-5", max_tokens=400,
+                                     messages=[{"role": "user", "content": prompt}])
+        return "".join(getattr(b, "text", "") for b in msg.content).strip()
+    if provider == "openai":
+        try:
+            import openai
+        except ImportError as exc:
+            raise ExplainError("Install the AI extra: pip install 'pipeverify[ai]'") from exc
+        if not model:
+            raise ExplainError("Set PIPEVERIFY_LLM_MODEL to the OpenAI model you want to use")
+        client = openai.OpenAI()
+        resp = client.chat.completions.create(model=model,
+                                              messages=[{"role": "user", "content": prompt}])
+        return (resp.choices[0].message.content or "").strip()
+    raise ExplainError(f"Unknown provider '{provider}'. Use anthropic or openai.")

pipeverify/report.py

@@ -0,0 +1,81 @@
+"""Render results for a terminal, a pull request comment (Markdown) or machines (JSON)."""
+from __future__ import annotations
+
+import json
+import os
+import sys
+
+from .core import RunResult
+
+ICON = {"pass": "✅", "review": "🟡", "block": "⛔", "info": "ℹ️"}
+LABEL = {"pass": "PASS", "review": "REVIEW", "block": "BLOCK", "info": "INFO"}
+HEADLINE = {
+    "pass": "Data looks the same where it should. Safe to merge.",
+    "review": "Something changed that tests would not catch. A person should look before merging.",
+    "block": "This change damages the data. Do not merge as is.",
+}
+
+
+def _headline(result: RunResult) -> str:
+    if any(t.error for t in result.tables):
+        return "Some tables could not be checked. Fix the errors below and run again."
+    return HEADLINE[result.verdict]
+
+
+def _color(text: str, sev: str) -> str:
+    if not sys.stdout.isatty() or os.environ.get("NO_COLOR"):
+        return text
+    code = {"pass": "32", "review": "33", "block": "31", "info": "36"}.get(sev, "0")
+    return f"\033[{code}m{text}\033[0m"
+
+
+def to_text(result: RunResult, show_info: bool = False) -> str:
+    lines = []
+    for t in result.tables:
+        lines.append(_color(f"{LABEL[t.verdict]:<7}", t.verdict) + f" {t.name}")
+        s = t.stats
+        if s:
+            extra = f", matched {s['matched_rows']:,}" if "matched_rows" in s else ""
+            lines.append(f"        rows {s.get('rows_before', 0):,} -> {s.get('rows_after', 0):,}{extra}")
+        for f in t.findings:
+            if f.severity == "info" and not show_info:
+                continue
+            lines.append("        " + _color(f"{LABEL[f.severity]:<6}", f.severity) + f" {f.message}")
+            ex = f.metrics.get("examples") if f.metrics else None
+            if ex:
+                lines.append(f"               e.g. {ex[0]}")
+    lines.append("")
+    lines.append(_color(f"Verdict: {LABEL[result.verdict]}", result.verdict) + f"  {_headline(result)}")
+    if result.explanation:
+        lines += ["", "Explanation:", result.explanation]
+    return "\n".join(lines)
+
+
+def to_markdown(result: RunResult) -> str:
+    out = [f"## {ICON[result.verdict]} pipeverify: {LABEL[result.verdict]}", "", _headline(result), ""]
+    out += ["| Table | Verdict | Rows before → after | Issues |", "|---|---|---|---|"]
+    for t in result.tables:
+        issues = sum(1 for f in t.findings if f.severity in ("review", "block"))
+        s = t.stats or {}
+        rows = f"{s.get('rows_before', '?'):,} → {s.get('rows_after', '?'):,}" if s else "n/a"
+        out.append(f"| `{t.name}` | {ICON[t.verdict]} {LABEL[t.verdict]} | {rows} | {issues} |")
+    for t in result.tables:
+        notable = [f for f in t.findings if f.severity in ("review", "block")]
+        info = [f for f in t.findings if f.severity == "info"]
+        if not notable and not info:
+            continue
+        out += ["", f"<details{' open' if notable else ''}><summary><b>{t.name}</b></summary>", ""]
+        for f in notable + info:
+            out.append(f"- {ICON[f.severity]} **{LABEL[f.severity]}** {f.message}")
+            ex = f.metrics.get("examples") if f.metrics else None
+            if ex:
+                out.append(f"  - example (key…, before, after): `{ex[0]}`")
+        out += ["", "</details>"]
+    if result.explanation:
+        out += ["", "### Explanation", "", result.explanation]
+    out += ["", f"<sub>Checked in {result.seconds:.2f}s by pipeverify. Verdicts compare data, not job status.</sub>"]
+    return "\n".join(out) + "\n"
+
+
+def to_json(result: RunResult) -> str:
+    return json.dumps(result.to_dict(), indent=2, default=str)

pipeverify/sources.py

@@ -0,0 +1,148 @@
+"""Load a table snapshot from a file, a DuckDB database, any SQL database or a DataFrame.
+
+Every source ends up as a table inside one DuckDB connection, where the checks run.
+"""
+from __future__ import annotations
+
+import os
+from dataclasses import dataclass
+from typing import Any, Mapping
+
+import duckdb
+
+FILE_READERS = {
+    ".csv": "read_csv_auto",
+    ".tsv": "read_csv_auto",
+    ".parquet": "read_parquet",
+    ".pq": "read_parquet",
+    ".json": "read_json_auto",
+    ".jsonl": "read_json_auto",
+    ".ndjson": "read_json_auto",
+}
+
+
+class SourceError(ValueError):
+    """Raised when a source spec cannot be understood or loaded."""
+
+
+@dataclass
+class Source:
+    """A normalized description of where a table snapshot lives."""
+
+    kind: str  # file | duckdb | sql | dataframe
+    path: str | None = None
+    table: str | None = None
+    query: str | None = None
+    url: str | None = None
+    frame: Any = None
+
+    def describe(self) -> str:
+        if self.kind == "file":
+            return self.path or ""
+        if self.kind == "duckdb":
+            return f"duckdb:{self.path}:{self.table or 'query'}"
+        if self.kind == "sql":
+            return f"sql:{_redact(self.url or '')}:{self.table or 'query'}"
+        return "dataframe"
+
+
+def _redact(url: str) -> str:
+    # hide passwords in connection strings that end up in reports
+    if "@" in url and "://" in url:
+        scheme, rest = url.split("://", 1)
+        creds, host = rest.split("@", 1)
+        user = creds.split(":", 1)[0]
+        return f"{scheme}://{user}:***@{host}"
+    return url
+
+
+def parse_source(spec: Any, base_dir: str = ".") -> Source:
+    """Turn a CLI string, a config mapping or a DataFrame into a Source.
+
+    String forms:
+      data/orders.parquet               a file (csv, parquet, json)
+      duckdb:warehouse.db:gold.orders   a table in a DuckDB file
+      duckdb:warehouse.db:SELECT ...    a query against a DuckDB file
+    Mapping forms (config file):
+      {path: ...} | {duckdb: ..., table|query: ...} | {url: ..., table|query: ...}
+    """
+    if hasattr(spec, "columns") and hasattr(spec, "dtypes"):
+        return Source(kind="dataframe", frame=spec)
+    if isinstance(spec, str):
+        if spec.startswith("duckdb:"):
+            parts = spec.split(":", 2)
+            if len(parts) != 3:
+                raise SourceError("DuckDB sources look like duckdb:<file.db>:<table or query>")
+            _, path, target = parts
+            path = _resolve(path, base_dir)
+            if target.strip().lower().startswith(("select", "with")):
+                return Source(kind="duckdb", path=path, query=target)
+            return Source(kind="duckdb", path=path, table=target)
+        return Source(kind="file", path=_resolve(spec, base_dir))
+    if isinstance(spec, Mapping):
+        if "path" in spec:
+            return Source(kind="file", path=_resolve(str(spec["path"]), base_dir))
+        if "duckdb" in spec:
+            return Source(kind="duckdb", path=_resolve(str(spec["duckdb"]), base_dir),
+                          table=spec.get("table"), query=spec.get("query"))
+        if "url" in spec:
+            url = os.path.expandvars(str(spec["url"]))
+            return Source(kind="sql", url=url, table=spec.get("table"), query=spec.get("query"))
+    raise SourceError(f"Unrecognized source: {spec!r}")
+
+
+def _resolve(path: str, base_dir: str) -> str:
+    path = os.path.expanduser(os.path.expandvars(path))
+    return path if os.path.isabs(path) else os.path.normpath(os.path.join(base_dir, path))
+
+
+def load(con: duckdb.DuckDBPyConnection, source: Source, name: str) -> None:
+    """Materialize a source as table `name` in the DuckDB connection."""
+    if source.kind == "dataframe":
+        con.register(f"{name}_df", source.frame)
+        con.execute(f'CREATE OR REPLACE TABLE "{name}" AS SELECT * FROM "{name}_df"')
+        con.unregister(f"{name}_df")
+        return
+    if source.kind == "file":
+        path = source.path or ""
+        if not os.path.exists(path):
+            raise SourceError(f"File not found: {path}")
+        ext = os.path.splitext(path)[1].lower()
+        reader = FILE_READERS.get(ext)
+        if reader is None:
+            raise SourceError(f"Unsupported file type '{ext}'. Use csv, parquet or json.")
+        con.execute(f'CREATE OR REPLACE TABLE "{name}" AS SELECT * FROM {reader}(?)', [path])
+        return
+    if source.kind == "duckdb":
+        if not source.path or not os.path.exists(source.path):
+            raise SourceError(f"DuckDB file not found: {source.path}")
+        alias = f"src_{name}"
+        lit = source.path.replace("'", "''")
+        con.execute(f"ATTACH '{lit}' AS {alias} (READ_ONLY)")
+        try:
+            if source.query:
+                con.execute(f"USE {alias}")
+                try:
+                    con.execute(f'CREATE OR REPLACE TABLE memory.main."{name}" AS {source.query}')
+                finally:
+                    con.execute("USE memory")
+            elif source.table:
+                con.execute(f'CREATE OR REPLACE TABLE "{name}" AS SELECT * FROM {alias}.{source.table}')
+            else:
+                raise SourceError("A DuckDB source needs a table or a query")
+        finally:
+            con.execute(f"DETACH {alias}")
+        return
+    if source.kind == "sql":
+        try:
+            import pandas as pd
+            import sqlalchemy
+        except ImportError as exc:  # pragma: no cover - depends on extras
+            raise SourceError("SQL sources need the sql extra: pip install 'pipeverify[sql]'") from exc
+        engine = sqlalchemy.create_engine(source.url)
+        query = source.query or f"SELECT * FROM {source.table}"
+        with engine.connect() as conn:
+            frame = pd.read_sql(sqlalchemy.text(query), conn)
+        load(con, Source(kind="dataframe", frame=frame), name)
+        return
+    raise SourceError(f"Unknown source kind {source.kind}")

pipeverify-0.1.0.dist-info/METADATA

@@ -0,0 +1,234 @@
+Metadata-Version: 2.5
+Name: pipeverify
+Version: 0.1.0
+Summary: Check that a pipeline change left the data right, not just the job green.
+Project-URL: Homepage, https://github.com/akshaysharma1088/pipeverify
+Project-URL: Issues, https://github.com/akshaysharma1088/pipeverify/issues
+Author: Akshay Sharma
+License-Expression: Apache-2.0
+License-File: LICENSE
+License-File: NOTICE
+Keywords: ci,data-diff,data-engineering,data-quality,duckdb,etl,llm-agents
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Database
+Classifier: Topic :: Software Development :: Quality Assurance
+Requires-Python: >=3.9
+Requires-Dist: click>=8.1
+Requires-Dist: duckdb>=1.0
+Requires-Dist: pyyaml>=6.0
+Provides-Extra: ai
+Requires-Dist: anthropic>=0.40; extra == 'ai'
+Requires-Dist: openai>=1.40; extra == 'ai'
+Provides-Extra: dev
+Requires-Dist: pandas>=2.0; extra == 'dev'
+Requires-Dist: pytest>=8; extra == 'dev'
+Provides-Extra: sql
+Requires-Dist: pandas>=2.0; extra == 'sql'
+Requires-Dist: sqlalchemy>=2.0; extra == 'sql'
+Description-Content-Type: text/markdown
+
+# pipeverify
+
+**Check that a pipeline change left the data right, not just the job green.**
+
+[![ci](https://github.com/akshaysharma1088/pipeverify/actions/workflows/ci.yml/badge.svg)](https://github.com/akshaysharma1088/pipeverify/actions/workflows/ci.yml)
+[![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](LICENSE)
+![Python](https://img.shields.io/badge/python-3.9%2B-blue)
+![Status](https://img.shields.io/badge/status-alpha-orange)
+
+A pipeline change can pass every test, finish on time and keep the schema intact while quietly
+dropping rows, duplicating keys or changing what a column means. More of these changes are now
+written by AI coding agents, which makes "the job succeeded" an even weaker signal.
+
+pipeverify compares your tables **before and after** a change and gives a verdict:
+
+| Verdict | Meaning | Default CI behaviour |
+|---|---|---|
+| ✅ **PASS** | The data is the same where it should be. | Merge |
+| 🟡 **REVIEW** | Something changed that tests would not catch. A person should look. | Merge allowed, flagged |
+| ⛔ **BLOCK** | The change damages the data. | Fail the check |
+
+It works with any tool that produces tables: SQL scripts, Spark jobs, dbt, Airflow tasks or an AI
+agent's fix. It needs two snapshots, nothing else.
+
+## Why
+
+In the [PipeMend benchmark](https://github.com/akshaysharma1088/pipemend)
+([DOI 10.5281/zenodo.23510784](https://doi.org/10.5281/zenodo.23510784)), LLM agents repaired
+64 data pipeline incidents. Production-style data quality monitors approved **every** harmful fix
+the agents made. What caught them was comparing the data itself. pipeverify packages that check
+so any team can run it on every change.
+
+## Quick start
+
+```bash
+pip install pipeverify
+
+# compare two snapshots of one table
+pipeverify compare before/orders.parquet after/orders.parquet --key order_id
+```
+
+```
+BLOCK   orders.parquet
+        rows 1,200 -> 1,000, matched 1,000
+        BLOCK  Table lost 200 rows (16.67%)
+        BLOCK  200 key(s) present before are gone after (16.67%)
+        BLOCK  Total of 'revenue' moved 16.71% (1,383,131.25 -> 1,152,000.00)
+
+Verdict: BLOCK  This change damages the data. Do not merge as is.
+```
+
+Try the bundled demo, five realistic changes to a retail sales table:
+
+```bash
+git clone https://github.com/akshaysharma1088/pipeverify && cd pipeverify
+pip install -e .
+python examples/retail/make_demo.py
+for s in examples/retail/scenarios/*/; do pipeverify run -c "$s/pipeverify.yml"; done
+```
+
+| Scenario | What happened | Verdict |
+|---|---|---|
+| `1_safe_fix` | Revenue restated for late refunds on one day, as intended | ✅ PASS |
+| `2_dedupe_drops_rows` | Duplicates removed on the wrong columns, real days dropped | ⛔ BLOCK |
+| `3_join_fanout` | A join to a dimension with a duplicate row doubled one store | ⛔ BLOCK |
+| `4_unit_change` | Upstream started sending one region's revenue in cents | ⛔ BLOCK |
+| `5_null_creep` | A renamed source field left region empty for new stores | 🟡 REVIEW |
+
+Every one of these runs successfully and keeps the same schema.
+
+## Sources
+
+`before` and `after` can each be:
+
+| Source | CLI form | Config form |
+|---|---|---|
+| CSV, Parquet, JSON file | `data/orders.parquet` | `path: data/orders.parquet` |
+| DuckDB table | `duckdb:warehouse.db:gold.orders` | `{duckdb: warehouse.db, table: gold.orders}` |
+| DuckDB query | `"duckdb:warehouse.db:SELECT ..."` | `{duckdb: warehouse.db, query: "SELECT ..."}` |
+| Any SQLAlchemy database (Postgres, Snowflake, BigQuery, ...) | | `{url: "postgresql://${PGUSER}@host/db", table: gold.orders}` |
+| pandas DataFrame | | Python API |
+
+SQL databases need `pip install "pipeverify[sql]"` plus the driver for your database. Environment
+variables in URLs are expanded, so credentials stay out of the config file.
+
+## Config
+
+Check many tables at once with `pipeverify.yml` (`pipeverify init` writes a starter file):
+
+```yaml
+thresholds:              # optional, overrides the defaults below for every table
+  row_change_block_pct: 10
+tables:
+  - name: gold.daily_sales
+    before: snapshots/main/daily_sales.parquet
+    after: snapshots/pr/daily_sales.parquet
+    key: [store_id, sale_date]
+    expect_change: [revenue]   # columns this change is meant to modify
+    ignore: [loaded_at]        # columns that always change
+  - name: gold.customers
+    before: {url: "${WAREHOUSE_URL}", table: prod.customers}
+    after: {url: "${WAREHOUSE_URL}", table: pr_123.customers}
+    key: [customer_id]
+    expect_row_change: true    # this change is meant to add or remove rows
+```
+
+```bash
+pipeverify run                     # uses ./pipeverify.yml
+pipeverify run --fail-on review    # stricter: fail CI on REVIEW too
+pipeverify run --format markdown -o report.md
+pipeverify run --json-out report.json
+```
+
+Exit code is `1` when the verdict reaches `--fail-on` (default `block`), otherwise `0`.
+
+## What it checks
+
+| Check | Catches | Default |
+|---|---|---|
+| Schema | Removed, added or retyped columns | removed: BLOCK, retyped: REVIEW |
+| Row count | Rows lost or gained | REVIEW > 1%, BLOCK ≥ 10% |
+| Missing keys | Rows that existed before and are gone | REVIEW on any, BLOCK ≥ 5% |
+| New keys | Rows that did not exist before | REVIEW |
+| Duplicate keys | Fan-out from joins, double loads | BLOCK on any new duplicate |
+| Null rate | Columns going empty | REVIEW > 1 pt, BLOCK ≥ 10 pts |
+| Changed values | Matched rows whose values changed, per column | REVIEW on any, BLOCK ≥ 10% |
+| Column totals | Unit changes, partial loads, double counting | REVIEW > 0.5%, BLOCK ≥ 5% |
+| Category values | New codes in low-cardinality text columns | REVIEW |
+
+Columns listed in `expect_change` are reported but never raise the verdict. All thresholds can be
+changed globally or per table; see [`pipeverify/checks.py`](pipeverify/checks.py) for every option.
+
+## GitHub Action
+
+```yaml
+# .github/workflows/pipeverify.yml
+on: pull_request
+permissions:
+  contents: read
+  pull-requests: write
+jobs:
+  verify:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      # ... build the "after" snapshots for this PR here ...
+      - uses: akshaysharma1088/pipeverify@v0.1.0
+        with:
+          config: pipeverify.yml
+          fail-on: block      # or review, or never
+```
+
+The action posts the report as a PR comment and in the job summary, and exposes the verdict as an
+output.
+
+## Python API
+
+```python
+import pipeverify
+
+result = pipeverify.verify(before_df, after_df, key="order_id", expect_change=["status"])
+print(result.verdict)              # "pass" | "review" | "block"
+for f in result.findings:
+    print(f.severity, f.message)
+```
+
+Useful inside an Airflow task, a notebook, or an AI agent's own loop: let the agent propose a fix,
+run `verify`, and only apply the fix on PASS.
+
+## Optional AI explanation
+
+```bash
+export ANTHROPIC_API_KEY=...       # or OPENAI_API_KEY with PIPEVERIFY_LLM_MODEL
+pip install "pipeverify[ai]"
+pipeverify run --explain
+```
+
+This adds a short plain-English explanation of what probably happened and what to check next.
+It is off by default. Only the findings are sent (check names, counts, percentages, column names);
+example values from your data are removed first. The verdict never depends on the LLM.
+
+## Roadmap
+
+- Native adapters for dbt (compare a PR's models against production), Airflow and Spark
+- Sampling and pushdown for very large tables, so checks run in the warehouse
+- Rules learned from history: what "normal" change looks like for each table
+- Agent integration: a guardrail that coding and data agents call before applying a change
+
+Ideas and pull requests are welcome; see [CONTRIBUTING.md](CONTRIBUTING.md).
+
+## Status
+
+Alpha. The checks and verdicts are stable enough to try in CI; options and output formats may
+still change before 1.0.
+
+## Citation
+
+If you use pipeverify in research, please cite the PipeMend benchmark that motivated it
+(see [CITATION.cff](CITATION.cff)).
+
+## License
+
+Apache-2.0

pipeverify-0.1.0.dist-info/RECORD

@@ -0,0 +1,13 @@
+pipeverify/__init__.py,sha256=0547yOB-etlZB5UyIrvI-OV2ZKnQaRz4UCt_GC8_TbM,349
+pipeverify/checks.py,sha256=aeqhFnPksVeGFo9oVMWpy0at1njrZAtLW5IoEcnLCDo,13087
+pipeverify/cli.py,sha256=rWh_UFOBVI5FcSevY0Lfx4eQaefTNvLVTxfg9CCccPU,6694
+pipeverify/core.py,sha256=7vYCMjU0-C6mWm8diZXLclJe7AtyQiLPw5VbNX2jjV8,4578
+pipeverify/explain.py,sha256=7tZBqFXkN5ySEtrNEt8dBXxm8dNYeFSSqb8GWxVLuOs,2950
+pipeverify/report.py,sha256=WBnDYA3mbIUOfMLcf8qGi8roykhh6HJfRWPUk2eaIjo,3619
+pipeverify/sources.py,sha256=XI2IfVaeYCD5cZ7be1kzbDHMaChq4G7mN87XxaXw3UM,5890
+pipeverify-0.1.0.dist-info/METADATA,sha256=Ldy55YLKZzgYFDD9oldW5sqFhIF9aq2o_3flVNdTeQE,9101
+pipeverify-0.1.0.dist-info/WHEEL,sha256=W3fkpkm7-wf9vBI5Z-7s0eWkeM-spu78I8Neb98DeEg,87
+pipeverify-0.1.0.dist-info/entry_points.txt,sha256=hg3tJswAuG4J0DJauqOQMrGZKjH7_5qA_C-mMQP7x-Y,51
+pipeverify-0.1.0.dist-info/licenses/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+pipeverify-0.1.0.dist-info/licenses/NOTICE,sha256=-Ei0_DtJMBByRJekSqwRIF-bZeDyj-zpkUHIXArp4KI,40
+pipeverify-0.1.0.dist-info/RECORD,,

pipeverify-0.1.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.32.4
+Root-Is-Purelib: true
+Tag: py3-none-any

pipeverify-0.1.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+pipeverify = pipeverify.cli:main

pipeverify-0.1.0.dist-info/licenses/LICENSE

@@ -0,0 +1,202 @@
+
+                                 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, whether in Source or
+      Object form, 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 modifications
+      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 Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to 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 discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated 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 combination of their Contribution(s)
+      with the Work to 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 a Contribution incorporated within 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, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, 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 as an addendum 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 copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   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 redistributing 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 consequential 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 any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional 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 act 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 such 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. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   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.

pipeverify-0.1.0.dist-info/licenses/NOTICE

@@ -0,0 +1,2 @@
+pipeverify
+Copyright 2026 Akshay Sharma