Support baidu vdb as doc engine #9984
Conversation
dosubot
bot
added
size:XL
letterbeezps
force-pushed
the
support_baidu_vdb
branch
from
0fe1e53 to
3691b3f
Compare
|
@KevinHuSh Hello, do you have time to review this code? BaiduVDB now supports field weight assignments. |
|
@letterbeezps Thanks for your contribution.
After these two issues can be resolved, we will merge it, thank you so much! |
| raise Exception(f"Mapping file not found at {fp_mapping}") | ||
| self.mapping = json.load(open(fp_mapping)) | ||
| healthy = self.health() | ||
| self.query_fields_boosts = { |
There was a problem hiding this comment.
@yingfeng the term weight is applied at here, This syntax is a little different from es, it acts directly on the query value rather than the field
There was a problem hiding this comment.
Is this field weight instead of term weight?
There was a problem hiding this comment.
Sorry, I'm a little confused. Can you point out where the term weight is used in "es_conn.py"！ so I can check and compare. @KevinHuSh
|
For the promotion of Baidu VDB，our team will maintain the baidu VDB as the doc engine for RAGFlow's new features |
| if syns and len(keywords) < 32: | ||
| keywords.extend(syns) | ||
| logging.debug(json.dumps(twts, ensure_ascii=False)) | ||
| for tk, w in sorted(twts, key=lambda x: x[1] * -1): |
There was a problem hiding this comment.
The w is the weight of the term, which has been dropped/neglected.
There was a problem hiding this comment.
OK, got it, I'll check it out. thanks
3 participants
What problem does this PR solve?
support use baidu vector database as doc engine
Type of change