gwc-pybundle 0.4.2__py3-none-any.whl

pybundle/steps/shell.py

@@ -0,0 +1,67 @@
+from __future__ import annotations
+
+import subprocess
+import time
+from dataclasses import dataclass
+from pathlib import Path
+
+from .base import StepResult
+from ..context import BundleContext
+
+
+@dataclass
+class ShellStep:
+    name: str
+    outfile_rel: str
+    cmd: list[str]
+    cwd_is_root: bool = True
+    allow_fail: bool = True
+    require_cmd: str | None = None
+
+    @property
+    def out_rel(self) -> str:
+        return self.outfile_rel
+
+    def run(self, ctx: BundleContext) -> StepResult:
+        if self.require_cmd and not getattr(ctx.tools, self.require_cmd, None):
+            out = ctx.workdir / self.outfile_rel
+            out.parent.mkdir(parents=True, exist_ok=True)
+            out.write_text(
+                f"{self.require_cmd} not found; skipping\n", encoding="utf-8"
+            )
+            return StepResult(self.name, "SKIP", 0, f"missing {self.require_cmd}")
+
+        out = ctx.workdir / self.outfile_rel
+        out.parent.mkdir(parents=True, exist_ok=True)
+
+        start = time.time()
+        header = (
+            f"## PWD: {ctx.root if self.cwd_is_root else Path.cwd()}\n"
+            f"## CMD: {' '.join(self.cmd)}\n\n"
+        )
+
+        try:
+            cp = subprocess.run(
+                self.cmd,
+                cwd=str(ctx.root) if self.cwd_is_root else None,
+                text=True,
+                capture_output=True,
+                check=False,
+            )
+            text = header + (cp.stdout or "") + ("\n" + cp.stderr if cp.stderr else "")
+            out.write_text(ctx.redact_text(text), encoding="utf-8")
+            status = (
+                "PASS"
+                if cp.returncode == 0
+                else ("FAIL" if not self.allow_fail else "PASS")
+            )
+            note = "" if cp.returncode == 0 else f"exit={cp.returncode}"
+        except Exception as e:
+            out.write_text(
+                ctx.redact_text(header + f"\nEXCEPTION: {e}\n"), encoding="utf-8"
+            )
+            status = "FAIL" if not self.allow_fail else "PASS"
+            note = str(e)
+
+        dur = int(time.time() - start)
+        return StepResult(self.name, status, dur, note)

pybundle/steps/tree.py

@@ -0,0 +1,136 @@
+from __future__ import annotations
+
+import os
+import time
+from dataclasses import dataclass, field
+from pathlib import Path
+
+from .base import StepResult
+from pybundle.context import BundleContext
+from pybundle.tools import which
+from pybundle.policy import AIContextPolicy, PathFilter
+
+BIN_EXTS = {
+    ".appimage", ".deb", ".rpm", ".exe", ".msi", ".dmg", ".pkg",
+    ".so", ".dll", ".dylib",
+}
+DB_EXTS = {".db", ".sqlite", ".sqlite3"}
+ARCHIVE_EXTS = {".zip", ".tar", ".gz", ".tgz", ".bz2", ".xz", ".7z"}
+
+DEFAULT_EXCLUDES = [
+    ".git",
+    ".venv",
+    ".mypy_cache",
+    ".ruff_cache",
+    ".pytest_cache",
+    "__pycache__",
+    "node_modules",
+    "dist",
+    "build",
+    "artifacts",
+    ".cache",
+]
+
+def _is_excluded(rel: Path, excludes: set[str]) -> bool:
+    # Exclude if any path component matches
+    for part in rel.parts:
+        if part in excludes:
+            return True
+    return False
+
+@dataclass
+class TreeStep:
+    name: str = "tree (filtered)"
+    max_depth: int = 4
+    excludes: list[str] | None = None
+    policy: AIContextPolicy | None = None
+
+    def run(self, ctx: BundleContext) -> StepResult:
+        start = time.time()
+        policy = self.policy or AIContextPolicy()
+
+        # allow overrides
+        exclude_dirs = set(self.excludes) if self.excludes else set(policy.exclude_dirs)
+        filt = PathFilter(exclude_dirs=exclude_dirs, exclude_file_exts=set(policy.exclude_file_exts))
+
+        out = ctx.metadir / "10_tree.txt"
+        out.parent.mkdir(parents=True, exist_ok=True)
+
+        root = ctx.root
+        lines: list[str] = []
+
+        for dirpath, dirnames, filenames in os.walk(root):
+            dp = Path(dirpath)
+            rel_dp = dp.relative_to(root)
+            depth = 0 if rel_dp == Path(".") else len(rel_dp.parts)
+
+            if depth > self.max_depth:
+                dirnames[:] = []
+                continue
+
+            # prune dirs (name + venv-structure)
+            kept = []
+            for d in dirnames:
+                if filt.should_prune_dir(dp, d):
+                    continue
+                kept.append(d)
+            dirnames[:] = kept
+
+            for fn in filenames:
+                p = dp / fn
+                if not filt.should_include_file(root, p):
+                    continue
+                lines.append(str(p.relative_to(root)))
+
+        lines.sort()
+        out.write_text("\n".join(lines) + ("\n" if lines else ""), encoding="utf-8")
+        dur = int(time.time() - start)
+        return StepResult(self.name, "PASS", dur, "python-walk")
+
+@dataclass
+class LargestFilesStep:
+    name: str = "largest files"
+    limit: int = 80
+    excludes: list[str] | None = None
+    policy: AIContextPolicy | None = None
+
+    def run(self, ctx: BundleContext) -> StepResult:
+        start = time.time()
+        policy = self.policy or AIContextPolicy()
+
+        exclude_dirs = set(self.excludes) if self.excludes else set(policy.exclude_dirs)
+        filt = PathFilter(exclude_dirs=exclude_dirs, exclude_file_exts=set(policy.exclude_file_exts))
+
+        out = ctx.metadir / "11_largest_files.txt"
+        out.parent.mkdir(parents=True, exist_ok=True)
+
+        files: list[tuple[int, str]] = []
+        root = ctx.root
+
+        for dirpath, dirnames, filenames in os.walk(root):
+            dp = Path(dirpath)
+
+            kept = []
+            for d in dirnames:
+                if filt.should_prune_dir(dp, d):
+                    continue
+                kept.append(d)
+            dirnames[:] = kept
+
+            for fn in filenames:
+                p = dp / fn
+                if not filt.should_include_file(root, p):
+                    continue
+                try:
+                    size = p.stat().st_size
+                except OSError:
+                    continue
+                files.append((size, str(p.relative_to(root))))
+
+        files.sort(key=lambda x: x[0], reverse=True)
+        lines = [f"{size}\t{path}" for size, path in files[: self.limit]]
+        out.write_text("\n".join(lines) + ("\n" if lines else ""), encoding="utf-8")
+
+        dur = int(time.time() - start)
+        return StepResult(self.name, "PASS", dur, f"count={len(files)}")
+

pybundle/tools.py

@@ -0,0 +1,7 @@
+from __future__ import annotations
+
+import shutil
+
+
+def which(cmd: str) -> str | None:
+    return shutil.which(cmd)