bincio-activity/bincio/render/merge.py

"""Apply sidecar .md edits to BAS JSON files.

Produces data_dir/_merged/ — a mirror of data_dir where:
- Files without sidecars are symlinked to the originals (cheap, preserves extracted data)
- Files with sidecars are written as merged copies
- index.json is rewritten with private filtering + highlight sort

This keeps data_dir/activities/*.json pristine (re-running extract never clobbers
user edits, and removing a sidecar always reverts fully on the next render).
"""

from __future__ import annotations

import json
import re
import shutil
from pathlib import Path

import yaml


def parse_sidecar(path: Path) -> tuple[dict, str]:
    """Return (frontmatter_dict, markdown_body) from a sidecar .md file."""
    text = path.read_text(encoding="utf-8")
    if text.startswith("---"):
        parts = re.split(r"^---[ \t]*$", text, maxsplit=2, flags=re.MULTILINE)
        if len(parts) >= 3:
            fm = yaml.safe_load(parts[1]) or {}
            return fm, parts[2].strip()
    return {}, text.strip()


def apply_sidecar(detail: dict, fm: dict, body: str) -> dict:
    """Apply sidecar overrides to a detail JSON dict. Returns a modified copy."""
    d = dict(detail)
    d.setdefault("custom", {})
    d["custom"] = dict(d["custom"])  # don't mutate original

    if "title" in fm:
        d["title"] = str(fm["title"])
    if "sport" in fm:
        d["sport"] = str(fm["sport"])
    if "gear" in fm:
        d["gear"] = str(fm["gear"]) if fm["gear"] else d.get("gear")
    if body:
        d["description"] = body
    elif "description" in fm:
        d["description"] = str(fm["description"])
    if "highlight" in fm:
        d["custom"]["highlight"] = bool(fm["highlight"])
    if "private" in fm:
        d["privacy"] = "private" if fm["private"] else detail.get("privacy", "public")
    if "hide_stats" in fm:
        d["custom"]["hide_stats"] = [str(s) for s in (fm["hide_stats"] or [])]

    return d


def _apply_sidecar_summary(summary: dict, fm: dict) -> dict:
    """Apply sidecar overrides to an index summary entry."""
    s = dict(summary)
    s.setdefault("custom", {})
    s["custom"] = dict(s["custom"])

    if "title" in fm:
        s["title"] = str(fm["title"])
    if "sport" in fm:
        s["sport"] = str(fm["sport"])
    if "highlight" in fm:
        s["custom"]["highlight"] = bool(fm["highlight"])
    if "private" in fm:
        s["privacy"] = "private" if fm["private"] else summary.get("privacy", "public")

    return s


def merge_all(data_dir: Path) -> int:
    """Build data_dir/_merged/ with all sidecar overrides applied.

    Returns the number of sidecars found and applied.
    """
    edits_dir = data_dir / "edits"
    acts_dir = data_dir / "activities"
    merged_dir = data_dir / "_merged"
    merged_acts = merged_dir / "activities"

    # Collect sidecars upfront
    sidecars: dict[str, tuple[dict, str]] = {}
    if edits_dir.exists():
        for md_path in sorted(edits_dir.glob("*.md")):
            sidecars[md_path.stem] = parse_sidecar(md_path)

    # Collect image lists — activities with uploaded images get custom.images even
    # if they have no sidecar text yet
    image_lists: dict[str, list[str]] = {}
    images_root = edits_dir / "images" if edits_dir.exists() else None
    if images_root and images_root.exists():
        for img_dir in sorted(images_root.iterdir()):
            if img_dir.is_dir():
                files = sorted(
                    p.name for p in img_dir.iterdir()
                    if p.is_file() and not p.name.startswith(".")
                )
                if files:
                    image_lists[img_dir.name] = files

    to_merge = set(sidecars) | set(image_lists)

    # Wipe and recreate _merged/activities/
    if merged_acts.exists():
        shutil.rmtree(merged_acts)
    merged_acts.mkdir(parents=True)

    # Mirror activities/ — symlink unmodified, write merged copies for overridden
    if acts_dir.exists():
        for src in sorted(acts_dir.iterdir()):
            if not src.is_file():
                continue
            dest = merged_acts / src.name
            activity_id = src.stem
            if src.suffix == ".json" and activity_id in to_merge:
                detail = json.loads(src.read_text(encoding="utf-8"))
                if activity_id in sidecars:
                    fm, body = sidecars[activity_id]
                    detail = apply_sidecar(detail, fm, body)
                if activity_id in image_lists:
                    detail["custom"] = dict(detail.get("custom") or {})
                    detail["custom"]["images"] = image_lists[activity_id]
                dest.write_text(json.dumps(detail, indent=2, ensure_ascii=False))
            else:
                dest.symlink_to(src.resolve())

    # Mirror edits/images/ → _merged/activities/images/ so the site can serve them
    if images_root and images_root.exists():
        merged_images = merged_acts / "images"
        merged_images.mkdir(exist_ok=True)
        for img_dir in images_root.iterdir():
            if img_dir.is_dir():
                dest_img = merged_images / img_dir.name
                if not dest_img.exists():
                    dest_img.symlink_to(img_dir.resolve())

    # Produce merged athlete.json — base from extract overlaid with edits/athlete.yaml
    athlete_src = data_dir / "athlete.json"
    athlete_dest = merged_dir / "athlete.json"
    if athlete_dest.exists() or athlete_dest.is_symlink():
        athlete_dest.unlink()
    if athlete_src.exists():
        athlete_edits_path = data_dir / "edits" / "athlete.yaml"
        if athlete_edits_path.exists():
            try:
                import yaml as _yaml
                edits = _yaml.safe_load(athlete_edits_path.read_text(encoding="utf-8")) or {}
            except Exception:
                edits = {}
        else:
            edits = {}
        if edits:
            athlete_data = json.loads(athlete_src.read_text(encoding="utf-8"))
            athlete_data.update(edits)
            athlete_dest.write_text(json.dumps(athlete_data, indent=2, ensure_ascii=False))
        else:
            athlete_dest.symlink_to(athlete_src.resolve())

    # Write merged index.json (private filtered, highlight sorted)
    index_path = data_dir / "index.json"
    if index_path.exists():
        index = json.loads(index_path.read_text(encoding="utf-8"))
        activities = []
        for s in index.get("activities", []):
            aid = s.get("id", "")
            if aid in sidecars:
                fm, _ = sidecars[aid]
                s = _apply_sidecar_summary(s, fm)
            activities.append(s)

        # Drop private activities from the published feed
        activities = [a for a in activities if a.get("privacy") != "private"]

        # Sort: newest first, then bring highlighted activities to the top
        activities.sort(key=lambda a: a.get("started_at", ""), reverse=True)
        activities.sort(key=lambda a: 0 if a.get("custom", {}).get("highlight") else 1)

        index["activities"] = activities
        (merged_dir / "index.json").write_text(
            json.dumps(index, indent=2, ensure_ascii=False)
        )
    elif (merged_dir / "index.json").exists():
        (merged_dir / "index.json").unlink()

    return len(sidecars)