#!/usr/bin/env bash
# Fetch PR task files into a workspace without git checkout.
# Usage: fetch_pr.sh <pr_number> <workspace_dir>
#
# Writes:
#   <workspace_dir>/pr-<n>-<task>/   — task files mirroring repo layout
#   <workspace_dir>/pr-<n>.meta.json — {pr, title, author, branch, head_repo, task_id}
#
# Requires: gh CLI authenticated to benchflow-ai/skillsbench.

set -euo pipefail

PR="${1:?pr number required}"
WS="${2:?workspace dir required}"
REPO="${SKILLSBENCH_REPO:-benchflow-ai/skillsbench}"

mkdir -p "$WS"

meta=$(gh pr view "$PR" --repo "$REPO" --json number,title,author,headRefName,headRepository,headRepositoryOwner)
branch=$(jq -r .headRefName <<<"$meta")
head_repo=$(jq -r '.headRepository.name' <<<"$meta")
head_owner=$(jq -r '.headRepositoryOwner.login' <<<"$meta")
author=$(jq -r '.author.login' <<<"$meta")
title=$(jq -r '.title' <<<"$meta")

files_json=$(gh api "repos/$REPO/pulls/$PR/files" --paginate)
task_files=$(jq -r '.[] | select(.filename | startswith("tasks/")) | .filename' <<<"$files_json")

if [[ -z "$task_files" ]]; then
    echo "PR #$PR has no task files" >&2
    exit 2
fi

task_id=$(awk -F/ 'NR==1{print $2}' <<<"$task_files")
target="$WS/pr-$PR-$task_id"
mkdir -p "$target"

while IFS= read -r f; do
    [[ -z "$f" ]] && continue
    dest="$target/$f"
    mkdir -p "$(dirname "$dest")"
    raw="https://raw.githubusercontent.com/$head_owner/$head_repo/$branch/$f"
    if curl -fsSL "$raw" -o "$dest"; then
        echo "  + $f"
    else
        echo "  ! failed: $f" >&2
    fi
done <<<"$task_files"

jq -n \
    --arg pr "$PR" --arg title "$title" --arg author "$author" \
    --arg branch "$branch" --arg head_repo "$head_owner/$head_repo" \
    --arg task_id "$task_id" --arg target "$target" \
    '{pr:$pr|tonumber, title:$title, author:$author, branch:$branch, head_repo:$head_repo, task_id:$task_id, target:$target}' \
    > "$WS/pr-$PR.meta.json"

echo "$target"