formatters.py

"""Turn an approved submission into a src/data/*.js entry + a text splice.

The site's data files (src/data/organizations.js, models.js, datasets.js,
blogs.js) are hand-formatted flat JS arrays, not JSON — see the real
examples this module's tests are based on. Rather than parsing/rewriting
the whole file as an AST, we do a plain text insertion, since the existing
formatting is simple and consistent: 2-space indent for `{`/`}`, 4-space
indent for fields, `description`/`excerpt` always on their own line.
"""

import json

# Canonical tag list — mirrors src/data/themes.js themeIds.
VALID_TAGS = [
    "biology", "chemistry", "physics", "medicine", "mathematics",
    "engineering", "earth-science", "astronomy", "genomics",
    "biotechnology", "materials-science", "climate", "energy",
    "ecology", "conservation", "benchmark", "scientific-reasoning",
]

# type -> (file path relative to repo root, exported array name, required fields)
TARGETS = {
    "organization": ("src/data/organizations.js", "organizations",
                      ["id", "name", "link", "tags"]),
    "model": ("src/data/models.js", "models",
              ["id", "slug", "name", "orgId", "type", "description", "tags"]),
    "dataset": ("src/data/datasets.js", "datasets",
                ["id", "slug", "orgId", "type", "description", "tags"]),
    # `slug` and `orgId` are legitimately absent for external (non-HF-blog)
    # and non-partnership posts — see e.g. the tamarind.bio entries in
    # blogs.js, most of which have slug: null and orgId: null.
    "blog": ("src/data/blogs.js", "blogs",
             ["id", "title", "date", "excerpt", "link", "tags"]),
}


def target_file(type_: str) -> str:
    return TARGETS[type_][0]


def missing_fields(type_: str, fields: dict) -> list[str]:
    _, _, required = TARGETS[type_]
    return [f for f in required if not fields.get(f)]


def _js_string(value) -> str:
    return json.dumps(value)


def render_entry(type_: str, fields: dict) -> str:
    """Render a single object literal matching the existing file style."""
    if type_ == "organization":
        lines = [
            "  {",
            f'    id: {_js_string(fields["id"])},',
            f'    name: {_js_string(fields["name"])},',
            f'    logo: getOrgLogo({_js_string(fields["id"])}),',
            "    description:",
            f'      {_js_string(fields.get("description", ""))},',
            f'    link: {_js_string(fields["link"])},',
            f'    tags: {json.dumps(fields["tags"])},',
            "  },",
        ]
    elif type_ == "model":
        lines = [
            "  {",
            f'    id: {_js_string(fields["id"])},',
            f'    slug: {_js_string(fields["slug"])},',
            f'    name: {_js_string(fields["name"])},',
            f'    orgId: {_js_string(fields["orgId"])},',
            f'    type: {_js_string(fields["type"])},',
            "    description:",
            f'      {_js_string(fields["description"])},',
            f'    tags: {json.dumps(fields["tags"])},',
            "  },",
        ]
    elif type_ == "dataset":
        lines = [
            "  {",
            f'    id: {_js_string(fields["id"])},',
            f'    slug: {_js_string(fields["slug"])},',
            f'    orgId: {_js_string(fields["orgId"])},',
            f'    type: {_js_string(fields["type"])},',
            "    description:",
            f'      {_js_string(fields["description"])},',
            f'    tags: {json.dumps(fields["tags"])},',
            "  },",
        ]
    elif type_ == "blog":
        lines = [
            "  {",
            f'    id: {_js_string(fields["id"])},',
            f'    title: {_js_string(fields["title"])},',
            f'    slug: {_js_string(fields["slug"]) if fields.get("slug") else "null"},',
            f'    orgId: {_js_string(fields["orgId"]) if fields.get("orgId") else "null"},',
            f'    date: {_js_string(fields["date"])},',
            "    excerpt:",
            f'      {_js_string(fields["excerpt"])},',
            f'    link: {_js_string(fields["link"])},',
            f'    tags: {json.dumps(fields["tags"])},',
            f'    featured: {"true" if fields.get("featured") else "false"},',
        ]
        if fields.get("upvotes") is not None:
            lines.append(f'    upvotes: {int(fields["upvotes"])},')
        lines.append("  },")
    else:
        raise ValueError(f"Unknown submission type: {type_}")

    return "\n".join(lines)


def insert_entry(file_content: str, type_: str, entry_block: str, org_id: str | None) -> str:
    """Insert entry_block into file_content, grouped near org_id's other
    entries when found, otherwise right after the array's opening line."""
    _, array_name, _ = TARGETS[type_]
    lines = file_content.split("\n")
    opening = f"export const {array_name} = ["

    insert_at = None
    if org_id:
        for i, line in enumerate(lines):
            if f'orgId: "{org_id}"' in line:
                # walk forward to this object's closing "  },"
                for j in range(i + 1, len(lines)):
                    if lines[j] == "  },":
                        insert_at = j + 1
                        break
                # keep scanning in case a later group is a better (last) match
    if insert_at is None:
        for i, line in enumerate(lines):
            if line.strip() == opening:
                insert_at = i + 1
                break

    if insert_at is None:
        raise ValueError(f"Could not find `{opening}` in target file")

    new_lines = lines[:insert_at] + entry_block.split("\n") + lines[insert_at:]
    return "\n".join(new_lines)