yumoqing/rag
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: use env.save_file/realpath instead of importing FileStorage in DSPY

Browse Source
This commit is contained in:
yumoqing 2026-07-30 10:38:06 +08:00
parent 3f3743842e
commit bd66e00a7a
2 changed files with 9 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
wwwroot/knowledge_bases_list/delete_file.dspy
View File

@ -29,12 +29,10 @@ async with get_sor_context(env, 'rag') as sor:
"UPDATE knowledge_bases SET doc_count=GREATEST(doc_count-1, 0), total_size=GREATEST(total_size-${size}$, 0) WHERE id=${kb_id}$", "UPDATE knowledge_bases SET doc_count=GREATEST(doc_count-1, 0), total_size=GREATEST(total_size-${size}$, 0) WHERE id=${kb_id}$",
{"size": file_size, "kb_id": kb_id}) {"size": file_size, "kb_id": kb_id})
# Delete physical file via FileStorage # Delete physical file via env.realpath
if file_path and file_path.startswith("/idfile/"): if file_path and file_path.startswith("/idfile/"):
from ahserver.filestorage import FileStorage
fs = FileStorage()
try: try:
real_path = fs.realPath(file_path) real_path = env.realpath(file_path)
if os.path.exists(real_path): if os.path.exists(real_path):
os.remove(real_path) os.remove(real_path)
except Exception: except Exception:

39
wwwroot/knowledge_bases_list/upload_file.dspy
View File

@ -12,15 +12,9 @@ if not file_data:
env = request._run_ns env = request._run_ns
userorgid = await env.get_userorgid() userorgid = await env.get_userorgid()
# Use FileStorage for proper hashed path # Save file via FileStorage (returns web path e.g. /idfile/191/193/197/97/xxx.txt)
from ahserver.filestorage import FileStorage web_path = await env.save_file(file_data, file_name)
fs = FileStorage() real_path = env.realpath(web_path)
real_path = fs._name2path(file_name)
with open(real_path, "wb") as f:
f.write(file_data)
web_path = fs.webpath(real_path) # e.g. /idfile/191/193/197/97/xxx.txt
if not web_path.startswith('/'):
web_path = '/' + web_path
doc_id = str(uuid()).replace('-', '')[:16] doc_id = str(uuid()).replace('-', '')[:16]
file_size = len(file_data) file_size = len(file_data)
@ -41,20 +35,6 @@ meta_parts = {}
# --- TEXT EXTRACTION --- # --- TEXT EXTRACTION ---
if ext_l in text_exts: if ext_l in text_exts:
text = file_data.decode('utf-8', errors='replace') text = file_data.decode('utf-8', errors='replace')
elif ext_l == '.pdf':
import io; from PyPDF2 import PdfReader
try:
r = PdfReader(io.BytesIO(file_data))
text = '\n'.join(p.extract_text() or '' for p in r.pages)
except: pass
elif ext_l == '.docx':
import io; from docx import Document
try:
doc = Document(io.BytesIO(file_data))
text = '\n'.join(p.text for p in doc.paragraphs if p.text.strip())
except: pass
import aiohttp, base64
# --- IMAGE: face detection --- # --- IMAGE: face detection ---
if ext_l in image_exts: if ext_l in image_exts:
@ -79,16 +59,12 @@ if ext_l in audio_exts:
r = await s.post('https://media.opencomputing.net/voiceprint/extract/submit', data=form) r = await s.post('https://media.opencomputing.net/voiceprint/extract/submit', data=form)
if r.status == 200: if r.status == 200:
vd = await r.json() vd = await r.json()
if vd.get('status') == 'SUCCEEDED': voice_speakers = vd.get('speakers', 1) if vd.get('status') == 'SUCCEEDED' else (1 if vd.get('embedding') else 0)
voice_speakers = vd.get('speakers', 1)
elif vd.get('embedding'):
voice_speakers = 1
meta_parts['voiceprint'] = voice_speakers meta_parts['voiceprint'] = voice_speakers
except: pass except: pass
# --- VIDEO: frame + audio extraction --- # --- VIDEO: frame extraction ---
if ext_l in video_exts: if ext_l in video_exts:
import subprocess
meta_parts['video'] = 'pending' meta_parts['video'] = 'pending'
try: try:
tmp_img = '/tmp/' + doc_id + '_frame.jpg' tmp_img = '/tmp/' + doc_id + '_frame.jpg'
@ -159,8 +135,7 @@ if text and len(text.strip()) > 10:
chunks_n = len(chunks) chunks_n = len(chunks)
# --- SAVE TO DB --- # --- SAVE TO DB ---
import json as _json meta_json = json.dumps(meta_parts, ensure_ascii=False)
meta_json = _json.dumps(meta_parts, ensure_ascii=False)
status = 'done' status = 'done'
async with get_sor_context(env, 'rag') as sor: async with get_sor_context(env, 'rag') as sor:
await sor.sqlExe( await sor.sqlExe(
@ -181,4 +156,4 @@ result = {"status": "SUCCEEDED", "doc_id": doc_id, "file_name": file_name,
"file_size": file_size, "folder_id": folder_id, "file_size": file_size, "folder_id": folder_id,
"text_len": len(text), "chunks": chunks_n, "text_len": len(text), "chunks": chunks_n,
"faces": face_count, "speakers": voice_speakers} "faces": face_count, "speakers": voice_speakers}
return _json.dumps(result, ensure_ascii=False, default=str) return json.dumps(result, ensure_ascii=False, default=str)