#!/usr/bin/env python3
"""Build the untracked-file triage manifest.

Universe: every file `git ls-files --others --exclude-standard` reports in
/root/troupe-standup (untracked, non-ignored). Each path is classified by the
rule in `category()` into exactly one bucket, and the result is written as a
strict TSV (path \t category \t rule_note) plus a JSON manifest.

Rule notes are verbatim, so the category of any row can be re-derived from the
path alone; the verifier does exactly that re-derivation.
"""
import json
import os
import subprocess
from collections import Counter

REPO = "/root/troupe-standup"
OUT_DIR = os.path.join(REPO, "wizard-work", "sol")
MANIFEST_TSV = os.path.join(OUT_DIR, "untracked-classify.tsv")
MANIFEST_JSON = os.path.join(OUT_DIR, "untracked-classify.json")
SUMMARY_MD = os.path.join(OUT_DIR, "untracked-classify-summary.md")

# Bucket order matters: earlier buckets win on overlap.
def category(p):
    base = os.path.basename(p)
    if p.startswith("wizard-work/null/"):
        return "probe"
    if p.startswith("wizard-work/spark/"):
        return "deliverable"
    if p.startswith("wizard-work/sphinx/"):
        return "report"
    if p.startswith("wizard-work/ledger/"):
        return "report"
    if p.startswith("wizard-work/"):
        return "other"
    if base.endswith(".py"):
        return "tool"
    if base.endswith((".json", ".csv", ".tsv", ".db", ".yaml", ".yml",
                      ".toml", ".lock", ".html", ".css")):
        return "data"
    if base.endswith((".md", ".txt")):
        return "report"
    return "other"

def rule_note(p):
    c = category(p)
    if c == "probe":
        return "wizard-work/null scratch probing output"
    if c == "deliverable":
        return "finished scripts + tests + output under wizard-work/spark"
    if c == "report":
        return "analysis / extraction / archive probe report"
    if c == "tool":
        return "standalone Python tool"
    if c == "data":
        return "data artifact (json/csv/tsv/db/yaml/toml/lock/html/css)"
    if c == "other":
        return "other file type"
    return "unclassified"

def main():
    out = subprocess.run(
        ["git", "-C", REPO, "ls-files", "--others", "--exclude-standard"],
        capture_output=True, text=True, check=True).stdout
    paths = sorted(set(out.splitlines()))
    paths = [p for p in paths if p]  # drop empties
    rows = [(p, category(p), rule_note(p)) for p in paths]

    os.makedirs(OUT_DIR, exist_ok=True)
    with open(MANIFEST_TSV, "w") as f:
        f.write("path\tcategory\trule_note\n")
        for p, c, n in rows:
            f.write(f"{p}\t{c}\t{n}\n")
    with open(MANIFEST_JSON, "w") as f:
        json.dump({"count": len(rows),
                   "snapshot_paths": paths,
                   "rows": rows,
                   "buckets": Counter(c for _, c, _ in rows)}, f, indent=2)
    with open(os.path.join(OUT_DIR, "untracked-classify-snapshot.txt"), "w") as f:
        f.write("\n".join(paths) + "\n")
    summary = ["# untracked-file triage manifest",
                "",
                "Universe: `git ls-files --others --exclude-standard` in the"
                " troupe repo, refreshed 2026-10-28.",
                "Vault (absolute path): " + MANIFEST_TSV,
                "Frozen universe (paths, one per line):"
                " " + os.path.join(OUT_DIR, "untracked-classify-snapshot.txt"),
                "Sum: %d files" % len(rows),
                ""] + ["%s: %d" % (k, v) for k, v in sorted(Counter(c for _, c, _ in rows).items())] + [""]
    open(SUMMARY_MD, "w").write("\n".join(summary))
    print("count", len(rows))
    for k, v in sorted(Counter(c for _, c, _ in rows).items()):
        print(k, v)

if __name__ == "__main__":
    main()