yangdx
ca2caf47bc
修改流式响应的输出格式：从event-stream改为x-ndjson
2025-01-15 22:14:57 +08:00
yangdx
6d44178f63
修复测试用例流结束判断
2025-01-15 21:26:20 +08:00
yangdx
af9ac188f0
增强聊天接口的调试和性能统计功能
...
- 添加原始请求日志记录
- 修改响应结构以包含性能统计
- 更新测试用例以展示性能数据
- 优化响应格式为字典结构
- 增加请求体解码功能
2025-01-15 21:15:12 +08:00
yangdx
8ef1248c76
将OllamaChatRequest的stream参数默认值改为True
2025-01-15 20:54:22 +08:00
yangdx
f81b1cdf0a
为Ollama API返回结果添加图像字段和性能统计信息
...
- 在OllamaMessage中添加images字段
- 响应消息中增加images字段
- 完成标记中添加性能统计信息
- 更新测试用例以处理性能统计
- 移除测试用例中的/naive前缀
2025-01-15 20:46:45 +08:00
yangdx
23f838ec94
优化流式响应处理并添加测试用例
...
- 修复流式响应中的完成标记逻辑
- 添加非流式调用测试
- 添加流式调用测试
- 优化JSON序列化，支持非ASCII字符
- 确保生成器在完成标记后立即结束
2025-01-15 20:18:17 +08:00
yangdx
f15f97a51d
临时保存
2025-01-15 19:32:03 +08:00
yangdx
828af49d6b
能够正确调用rag，rag执行完成后，无法返回内容
2025-01-15 18:47:01 +08:00
yangdx
4e5517a602
修复rag调用参数不正确问题
2025-01-15 18:27:35 +08:00
yangdx
882da88603
解决模型名称返回错误问题
2025-01-15 18:19:39 +08:00
yangdx
fd50c3a240
修改流处理逻辑，修改 /api/tags错误
2025-01-15 17:43:00 +08:00
yangdx
c1f4f4a20e
优化ollama返回数据的格式
2025-01-15 15:06:28 +08:00
yangdx
be134878fe
完成ollma接口的代码编写
2025-01-15 14:31:49 +08:00
yangdx
b97d1ecd72
修改文件名
2025-01-15 13:35:20 +08:00
yangdx
9bfba88600
准备增加ollama服务
2025-01-15 13:32:06 +08:00
yangdx
da91510363
修改server启动命令
2025-01-15 13:14:09 +08:00
yangdx
1088e10fb2
修改lightrag_server的LLM和Embedding配置
2025-01-15 10:44:12 +08:00
zrguo
61349b0fe7
Merge pull request #581 from jiabin2wang20230918/fix_mongo
...
Fix for mongoDB document 16MB limit and add os env NEO4J_MAX_CONNECTION_POOL_SIZE for neo4j
2025-01-13 18:38:22 +08:00
bingo
1984da0fd6
add logger.debug for mongo_impl get_by_mode_and_id()
2025-01-13 07:27:30 +00:00
bingo
f28b90b2b3
1. add os env NEO4J_MAX_CONNECTION_POOL_SIZE to for neo4j ;
...
2. fix https://github.com/HKUDS/LightRAG/issues/580 issue for mongoDB document 16MB limit.
2025-01-13 07:06:01 +00:00
zrguo
057e23c4e9
Update __init__.py
2025-01-13 10:13:01 +08:00
zrguo
9f48124a69
Merge pull request #576 from ShanGor/main
...
Revise the AGE implementation on get_node_edges, to align with Neo4j behavior.
2025-01-13 09:29:01 +08:00
Samuel Chan
c016934021
Revise the AGE implementation on get_node_edges, to align with Neo4j behavior.
2025-01-12 21:38:39 +08:00
Saifeddine ALOUI
7aaab219ee
Fixed awaiting insert
2025-01-12 12:56:08 +01:00
Saifeddine ALOUI
5c67938467
Resolve 500 error caused by missing len() for LightRAG's API insert_text endpoint
2025-01-12 12:46:23 +01:00
zrguo
a2e96b67e9
Merge pull request #570 from ShanGor/main
...
Revise the AGE usage for postgres_impl
2025-01-12 13:23:06 +08:00
Samuel Chan
1998a5b204
Merge remote-tracking branch 'origin/main'
...
# Conflicts:
# README.md
2025-01-11 10:40:09 +08:00
Samuel Chan
d03d6f5fc5
Revised the postgres implementation, to use attributes(node_id) rather than nodes to identify an entity. Which significantly reduced the table counts.
2025-01-11 09:30:19 +08:00
Saifeddine ALOUI
224fce9b1b
run precommit to fix linting issues
2025-01-11 01:37:07 +01:00
Saifeddine ALOUI
e0e656ab01
Added ssl support
2025-01-11 01:35:49 +01:00
Saifeddine ALOUI
a619b01064
Next test of timeout
2025-01-10 22:17:13 +01:00
Saifeddine ALOUI
ab3cc3f0f4
fixed missing coma
2025-01-10 21:39:41 +01:00
Saifeddine ALOUI
adb288c5bb
added timeout
2025-01-10 21:39:25 +01:00
Saifeddine ALOUI
2297007b7b
Simplified the api services issue #565
2025-01-10 20:30:58 +01:00
adikalra
acde4ed173
Add custom chunking function.
2025-01-09 17:20:24 +05:30
zrguo
b93203804c
Merge branch 'main' into main
2025-01-09 15:28:57 +08:00
zrguo
92ccfa2770
Merge pull request #555 from ParisNeo/main
...
Restore backwards compatibility for LightRAG's ainsert method
2025-01-09 15:27:09 +08:00
童石渊
dd213c95be
增加仅字符分割参数，如果开启，仅采用字符分割，不开启，在分割完以后如果chunk过大，会继续根据token size分割，更新测试文件
2025-01-09 11:55:49 +08:00
Saifeddine ALOUI
65c1450c66
fixed retro compatibility with ainsert by making split_by_character get a None default value
2025-01-08 20:50:22 +01:00
Gurjot Singh
9565a4663a
Fix trailing whitespace and formatting issues in lightrag.py
2025-01-09 00:39:22 +05:30
Gurjot Singh
a940251390
Implement custom chunking feature
2025-01-07 20:57:39 +05:30
zrguo
6c78c96854
fix linting errors
2025-01-07 22:02:34 +08:00
zrguo
fe7f7086b1
Merge pull request #547 from n3A87/main
...
Fix:Optimized logic for automatic switching modes when keywords do not exist
2025-01-07 21:51:51 +08:00
童石渊
6b19401dc6
chunk split retry
2025-01-07 16:26:12 +08:00
童石渊
536d6f2283
添加字符分割功能，在“insert”函数中如果增加参数split_by_character，则会按照split_by_character进行字符分割，此时如果每个分割后的chunk的tokens大于max_token_size，则会继续按token_size分割（todo：考虑字符分割后过短的chunk处理）
2025-01-07 00:28:15 +08:00
xYLiuuuuuu
79646fced8
Fix:Optimized logic for automatic switching modes when keywords do not exist
2025-01-06 16:54:53 +08:00
zrguo
39a366a3dc
Update __init__.py
2025-01-06 15:43:05 +08:00
zrguo
990b684a85
Update lightrag.py
2025-01-06 15:27:31 +08:00
Samuel Chan
85d35b16d4
Merge branch 'HKUDS:main' into main
2025-01-06 12:53:06 +08:00
Samuel Chan
6ae27d8f06
Some enhancements:
...
- Enable the llm_cache storage to support get_by_mode_and_id, to improve the performance for using real KV server
- Provide an option for the developers to cache the LLM response when extracting entities for a document. Solving the paint point that sometimes the process failed, the processed chunks we need to call LLM again, money and time wasted. With the new option (by default not enabled) enabling, we can cache that result, can significantly save the time and money for beginners.
2025-01-06 12:50:05 +08:00
