Download convert_SQLITE_index_stories.bashpy from tgdatasets/writing.com_database: direct link, hf CLI and curl.
- Browser
- Download file 4.16 kB
-
https://huggingface.co/datasets/tgdatasets/writing.com_database/resolve/main/convert_SQLITE_index_stories.bashpy
- Command line
-
hf download hf://datasets/tgdatasets/writing.com_database/convert_SQLITE_index_stories.bashpy
-
curl -L -o convert_SQLITE_index_stories.bashpy https://huggingface.co/datasets/tgdatasets/writing.com_database/resolve/main/convert_SQLITE_index_stories.bashpy
4.16 kB
| import argparse | |
| import json | |
| import os | |
| import sys | |
| import sqlite3 | |
| from typing import Dict, Any | |
| def main() -> int: | |
| ap = argparse.ArgumentParser( | |
| description="Build a JSON index from story-database.sqlite." | |
| ) | |
| ap.add_argument("--db", default="story-database.sqlite", help="Input SQLite database path") | |
| ap.add_argument( | |
| "--output", | |
| default="index.json", | |
| help="Output JSON path (default: index.json)", | |
| ) | |
| args = ap.parse_args() | |
| if not os.path.exists(args.db): | |
| print(f"Database not found: {args.db}") | |
| return 1 | |
| conn = sqlite3.connect(args.db) | |
| conn.row_factory = sqlite3.Row | |
| cursor = conn.cursor() | |
| out: Dict[str, Any] = {} | |
| # Read from stories table. Note that we will compute the correct chapter count | |
| # by joining with chapters, handling the Intro emptiness logic if needed. | |
| # However, since the sqlite database likely already has chapter_count computed | |
| # accurately or we can just fetch it, let's look at what we need. | |
| # We will get all stories. | |
| # The previous logic counted chapters and excluded "-Intro" if chapter_content was empty. | |
| # Since we're using SQLite, we can just fetch the chapter_count from stories, but | |
| # to be perfectly safe and replicate the logic, let's query chapters as well to adjust if necessary, | |
| # or just trust the stories table if it's already accurate. | |
| # The prompt says: "make this to work on story-database.sqlite (it has the same logic story-database.tsv)" | |
| # We'll calculate the chapter counts directly from chapters table for accuracy, ignoring empty intros. | |
| # Wait, fetching all chapters content might be heavy. Let's just use a query. | |
| print(f"Reading from {args.db}...") | |
| # Query to get story data | |
| cursor.execute("SELECT base_id, title, author_name, author_raw_name, synopsis, date, last_modified FROM stories") | |
| stories = cursor.fetchall() | |
| # Query to get chapter counts per base_id | |
| # We count all chapters, but we need to exclude "-Intro" if content is empty or {"html": ""} | |
| cursor.execute(''' | |
| SELECT base_id, map_id, chapter_content FROM chapters | |
| ''') | |
| chapter_count = {} | |
| for row in cursor: | |
| base_id = row['base_id'] | |
| map_id = row['map_id'] | |
| content = (row['chapter_content'] or "").strip() | |
| is_empty_intro = False | |
| if "-Intro" in map_id: | |
| if not content or content in ('{"html": ""}', '{"html":""}'): | |
| is_empty_intro = True | |
| elif content.startswith("{") and content.endswith("}"): | |
| if '"html": ""' in content or '"html":""' in content: | |
| try: | |
| obj = json.loads(content) | |
| if isinstance(obj, dict) and obj.get("html", "").strip() == "": | |
| is_empty_intro = True | |
| except: | |
| pass | |
| if not is_empty_intro: | |
| chapter_count[base_id] = chapter_count.get(base_id, 0) + 1 | |
| for row in stories: | |
| base_id = row['base_id'] | |
| if not base_id or base_id.startswith('#====') or (row['title'] and row['title'].startswith('#====')): | |
| continue | |
| c = chapter_count.get(base_id, 0) | |
| out[base_id] = { | |
| "title": row["title"] or "", | |
| "authorcutename": row["author_name"] or "", | |
| "authorid": row["author_raw_name"] or "", | |
| "synopsis": row["synopsis"] or "", | |
| "date": row["date"] or "", | |
| "creationdate": row["date"] or "", | |
| "lastmodified": row["last_modified"] or "", | |
| "chapterCount": c, | |
| "isonline": "", | |
| } | |
| conn.close() | |
| out = {k: out[k] for k in sorted(out)} | |
| with open(args.output, "w", encoding="utf-8", newline="\n") as f: | |
| json.dump( | |
| out, | |
| f, | |
| ensure_ascii=False, | |
| indent=2, | |
| separators=(",", ": "), | |
| ) | |
| f.write("\n") | |
| print(f"Saved {len(out)} stories to {args.output}") | |
| return 0 | |
| if __name__ == "__main__": | |
| sys.exit(main()) | |