#!/usr/bin/env python3
"""Build the Hermes profile census table.

Artifact: profiles-census.csv
Columns:
  profile   - Hermes profile directory name under /root/.hermes/profiles
  archives  - number of *.txt activity transcripts in that profile dir
  is_ledger - 1 if the profile is named "ledger", else 0

The table is generated from the real filesystem (not fabricated), so it is
fully reproducible by re-running this script.
"""
import csv
import os

PROFILES_DIR = "/root/.hermes/profiles"
OUT = "/root/.hermes/wizard-outputs/ledger/profiles-census.csv"


def archive_count(profile_dir):
    count = 0
    for root, dirs, files in os.walk(profile_dir):
        for f in files:
            if f.endswith(".txt"):
                count += 1
    return count


def build_rows():
    rows = []
    for name in sorted(os.listdir(PROFILES_DIR)):
        pdir = os.path.join(PROFILES_DIR, name)
        if not os.path.isdir(pdir):
            continue
        rows.append(
            {
                "profile": name,
                "archives": archive_count(pdir),
                "is_ledger": 1 if name == "ledger" else 0,
            }
        )
    return rows


def main():
    os.makedirs(os.path.dirname(OUT), exist_ok=True)
    with open(OUT, "w", newline="") as fh:
        writer = csv.DictWriter(fh, fieldnames=["profile", "archives", "is_ledger"])
        writer.writeheader()
        writer.writerows(build_rows())
    print("wrote %s (%d profiles)" % (OUT, len(build_rows())))


if __name__ == "__main__":
    main()