#!/usr/bin/env python3
"""
Gitea Language Stats
Calculates the percentage of every programming language used
in all repositories of a given account and prints a markdown
table to stdout, delimited by the
markers  <!--LANG_STATS_START--> … <!--LANG_STATS_END-->.

Required environment variables
--------------------------------
GITEA_BASE_URL   e.g. https://gitea.example.com
GITEA_TOKEN      personal access-token with `repo` scope
GITEA_USERNAME   the account to analyse
"""
import os
import sys
import requests
from collections import defaultdict
from dotenv import load_dotenv

load_dotenv()

BASE_URL   = os.getenv("GITEA_BASE_URL", "").rstrip("/")
TOKEN      = os.getenv("GITEA_TOKEN")
USERNAME   = os.getenv("GITEA_USERNAME")

if not BASE_URL or not TOKEN or not USERNAME:
    sys.exit("GITEA_BASE_URL, GITEA_TOKEN and GITEA_USERNAME must be set")

session = requests.Session()
session.headers["Authorization"] = f"token {TOKEN}"

# List of files whose byte-counts should be excluded from the totals.
# Each tuple: (repository_name, file_path, language_name)
EXCLUDE_FILES: list[tuple[str, str, str]] = [
    ("ar-scavenger-hunt", "static/aframe.js", "JavaScript"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.4.html", "HTML"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.4-css.html", "HTML"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.5.html", "HTML"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.5-css.html", "HTML"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.6.html", "HTML"),
    ("Kingsborough-LibGuide", "groups/home/bootstrap3/Vue3-v.6-css.html", "HTML"),
]

# ---------- collect repositories ----------
page = 1
repos = []
while True:
    resp = session.get(
        f"{BASE_URL}/api/v1/users/{USERNAME}/repos",
        params={"page": page, "limit": 50},
        timeout=30,
    )
    resp.raise_for_status()
    batch = resp.json()
    if not batch:
        break
    repos.extend(batch)
    page += 1

# ---------- aggregate language bytes ----------
totals: dict[str, int] = defaultdict(int)

for repo in repos:
    langs = session.get(
        f"{BASE_URL}/api/v1/repos/{USERNAME}/{repo['name']}/languages",
        timeout=30,
    ).json()
    for lang, bytes_ in langs.items():
        totals[lang] += bytes_

# ---------- subtract excluded files ----------
for repo_name, file_path, language in EXCLUDE_FILES:
    try:
        meta = session.get(
            f"{BASE_URL}/api/v1/repos/{USERNAME}/{repo_name}/contents/{file_path}",
            timeout=30,
        )
        if meta.ok:
            size = meta.json().get("size", 0)
            totals[language] = max(0, totals[language] - size)
    except requests.RequestException:
        # Ignore failures – continue with best-effort stats
        pass

total_bytes = sum(totals.values())
if total_bytes == 0:
    sys.exit("No language data returned from API.")

# ---------- build markdown table ----------
pairs = sorted(totals.items(), key=lambda x: x[1], reverse=True)

lines = [
    "## 📊 Languages\n",
]
for lang, nbytes in pairs:
    pct = nbytes * 100 / total_bytes
    bar = "█" * int(pct / 2)  # up to 50 chars
    lines.append(f"{bar} {lang}{pct:5.1f}%\\")

for line in lines:
    print(line)