dataforge
c9fd4aa84c
- Añade enricher.go + directorio enrichers/ para enriquecer entidades con fuentes externas. - Nuevos componentes frontend: IngestPanel (panel de ingesta de datos) y NodeContextMenu (menu contextual sobre nodos del grafo). - Retira SearchBar y lib/utils.ts; la busqueda se integra dentro de los paneles existentes. - Ajusta tipos (types.go, types.ts, wailsjs/go) y theming (postcss + app.css + Mantine). - Actualiza app.go y wails.json para exponer las nuevas capacidades. - Añade directorio projects/ con estado inicial. - Rebuild del frontend (dist actualizado).
59 lines
1.6 KiB
Python
59 lines
1.6 KiB
Python
"""Enricher: Extract text from a document file."""
|
|
|
|
import sys
|
|
import json
|
|
import os
|
|
|
|
sys.path.insert(0, os.path.join(os.environ.get("FN_REGISTRY_ROOT", ""), "python", "functions", "core"))
|
|
|
|
from extract_text_from_file import extract_text_from_file
|
|
|
|
|
|
def main():
|
|
entity = json.load(sys.stdin)
|
|
file_path = (entity.get("metadata") or {}).get("file_path", "")
|
|
|
|
if not file_path:
|
|
json.dump({"error": "No file_path in entity metadata"}, sys.stdout)
|
|
return
|
|
|
|
if not os.path.exists(file_path):
|
|
json.dump({"error": f"File not found: {file_path}"}, sys.stdout)
|
|
return
|
|
|
|
text = extract_text_from_file(file_path)
|
|
|
|
result = {
|
|
"entities": [
|
|
{
|
|
"name": f"Text: {os.path.basename(file_path)}",
|
|
"type_ref": "text",
|
|
"description": f"Text extracted from {os.path.basename(file_path)}",
|
|
"tags": ["extracted"],
|
|
"metadata": {
|
|
"content_preview": text[:500],
|
|
"source": file_path,
|
|
"char_count": len(text),
|
|
"full_content": text,
|
|
},
|
|
"notes": "",
|
|
}
|
|
],
|
|
"relations": [
|
|
{
|
|
"name": "extracted_from",
|
|
"from_entity": "__NEW_0__",
|
|
"to_entity": "__SOURCE__",
|
|
"description": f"Text extracted from document",
|
|
"weight": 1.0,
|
|
"tags": [],
|
|
"notes": "",
|
|
}
|
|
],
|
|
}
|
|
json.dump(result, sys.stdout, ensure_ascii=False)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|