Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
chenpangpang
open-webui
Commits
9755cd5b
"...resnet50_tensorflow.git" did not exist on "2912042352009c9993dc05403624100bfe42d9c1"
Commit
9755cd5b
authored
Apr 25, 2024
by
Steven Kreitzer
Browse files
feat: toggle hybrid search
parent
984dbf13
Changes
4
Hide whitespace changes
Inline
Side-by-side
Showing
4 changed files
with
132 additions
and
87 deletions
+132
-87
backend/apps/rag/main.py
backend/apps/rag/main.py
+17
-1
backend/apps/rag/utils.py
backend/apps/rag/utils.py
+8
-3
backend/config.py
backend/config.py
+1
-0
src/lib/components/documents/Settings/General.svelte
src/lib/components/documents/Settings/General.svelte
+106
-83
No files found.
backend/apps/rag/main.py
View file @
9755cd5b
...
@@ -70,6 +70,7 @@ from config import (
...
@@ -70,6 +70,7 @@ from config import (
RAG_EMBEDDING_MODEL
,
RAG_EMBEDDING_MODEL
,
RAG_EMBEDDING_MODEL_AUTO_UPDATE
,
RAG_EMBEDDING_MODEL_AUTO_UPDATE
,
RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE
,
RAG_EMBEDDING_MODEL_TRUST_REMOTE_CODE
,
RAG_HYBRID
,
RAG_RERANKING_MODEL
,
RAG_RERANKING_MODEL
,
RAG_RERANKING_MODEL_AUTO_UPDATE
,
RAG_RERANKING_MODEL_AUTO_UPDATE
,
RAG_RERANKING_MODEL_TRUST_REMOTE_CODE
,
RAG_RERANKING_MODEL_TRUST_REMOTE_CODE
,
...
@@ -91,6 +92,8 @@ app = FastAPI()
...
@@ -91,6 +92,8 @@ app = FastAPI()
app
.
state
.
TOP_K
=
RAG_TOP_K
app
.
state
.
TOP_K
=
RAG_TOP_K
app
.
state
.
RELEVANCE_THRESHOLD
=
RAG_RELEVANCE_THRESHOLD
app
.
state
.
RELEVANCE_THRESHOLD
=
RAG_RELEVANCE_THRESHOLD
app
.
state
.
HYBRID
=
RAG_HYBRID
app
.
state
.
CHUNK_SIZE
=
CHUNK_SIZE
app
.
state
.
CHUNK_SIZE
=
CHUNK_SIZE
app
.
state
.
CHUNK_OVERLAP
=
CHUNK_OVERLAP
app
.
state
.
CHUNK_OVERLAP
=
CHUNK_OVERLAP
...
@@ -321,6 +324,7 @@ async def get_query_settings(user=Depends(get_admin_user)):
...
@@ -321,6 +324,7 @@ async def get_query_settings(user=Depends(get_admin_user)):
"template"
:
app
.
state
.
RAG_TEMPLATE
,
"template"
:
app
.
state
.
RAG_TEMPLATE
,
"k"
:
app
.
state
.
TOP_K
,
"k"
:
app
.
state
.
TOP_K
,
"r"
:
app
.
state
.
RELEVANCE_THRESHOLD
,
"r"
:
app
.
state
.
RELEVANCE_THRESHOLD
,
"hybrid"
:
app
.
state
.
HYBRID
,
}
}
...
@@ -328,6 +332,7 @@ class QuerySettingsForm(BaseModel):
...
@@ -328,6 +332,7 @@ class QuerySettingsForm(BaseModel):
k
:
Optional
[
int
]
=
None
k
:
Optional
[
int
]
=
None
r
:
Optional
[
float
]
=
None
r
:
Optional
[
float
]
=
None
template
:
Optional
[
str
]
=
None
template
:
Optional
[
str
]
=
None
hybrid
:
Optional
[
bool
]
=
None
@
app
.
post
(
"/query/settings/update"
)
@
app
.
post
(
"/query/settings/update"
)
...
@@ -337,7 +342,14 @@ async def update_query_settings(
...
@@ -337,7 +342,14 @@ async def update_query_settings(
app
.
state
.
RAG_TEMPLATE
=
form_data
.
template
if
form_data
.
template
else
RAG_TEMPLATE
app
.
state
.
RAG_TEMPLATE
=
form_data
.
template
if
form_data
.
template
else
RAG_TEMPLATE
app
.
state
.
TOP_K
=
form_data
.
k
if
form_data
.
k
else
4
app
.
state
.
TOP_K
=
form_data
.
k
if
form_data
.
k
else
4
app
.
state
.
RELEVANCE_THRESHOLD
=
form_data
.
r
if
form_data
.
r
else
0.0
app
.
state
.
RELEVANCE_THRESHOLD
=
form_data
.
r
if
form_data
.
r
else
0.0
return
{
"status"
:
True
,
"template"
:
app
.
state
.
RAG_TEMPLATE
}
app
.
state
.
HYBRID
=
form_data
.
hybrid
if
form_data
.
hybrid
else
False
return
{
"status"
:
True
,
"template"
:
app
.
state
.
RAG_TEMPLATE
,
"k"
:
app
.
state
.
TOP_K
,
"r"
:
app
.
state
.
RELEVANCE_THRESHOLD
,
"hybrid"
:
app
.
state
.
HYBRID
,
}
class
QueryDocForm
(
BaseModel
):
class
QueryDocForm
(
BaseModel
):
...
@@ -345,6 +357,7 @@ class QueryDocForm(BaseModel):
...
@@ -345,6 +357,7 @@ class QueryDocForm(BaseModel):
query
:
str
query
:
str
k
:
Optional
[
int
]
=
None
k
:
Optional
[
int
]
=
None
r
:
Optional
[
float
]
=
None
r
:
Optional
[
float
]
=
None
hybrid
:
Optional
[
bool
]
=
None
@
app
.
post
(
"/query/doc"
)
@
app
.
post
(
"/query/doc"
)
...
@@ -368,6 +381,7 @@ def query_doc_handler(
...
@@ -368,6 +381,7 @@ def query_doc_handler(
r
=
form_data
.
r
if
form_data
.
r
else
app
.
state
.
RELEVANCE_THRESHOLD
,
r
=
form_data
.
r
if
form_data
.
r
else
app
.
state
.
RELEVANCE_THRESHOLD
,
embeddings_function
=
embeddings_function
,
embeddings_function
=
embeddings_function
,
reranking_function
=
app
.
state
.
sentence_transformer_rf
,
reranking_function
=
app
.
state
.
sentence_transformer_rf
,
hybrid
=
form_data
.
hybrid
if
form_data
.
hybrid
else
app
.
state
.
HYBRID
,
)
)
except
Exception
as
e
:
except
Exception
as
e
:
log
.
exception
(
e
)
log
.
exception
(
e
)
...
@@ -382,6 +396,7 @@ class QueryCollectionsForm(BaseModel):
...
@@ -382,6 +396,7 @@ class QueryCollectionsForm(BaseModel):
query
:
str
query
:
str
k
:
Optional
[
int
]
=
None
k
:
Optional
[
int
]
=
None
r
:
Optional
[
float
]
=
None
r
:
Optional
[
float
]
=
None
hybrid
:
Optional
[
bool
]
=
None
@
app
.
post
(
"/query/collection"
)
@
app
.
post
(
"/query/collection"
)
...
@@ -405,6 +420,7 @@ def query_collection_handler(
...
@@ -405,6 +420,7 @@ def query_collection_handler(
r
=
form_data
.
r
if
form_data
.
r
else
app
.
state
.
RELEVANCE_THRESHOLD
,
r
=
form_data
.
r
if
form_data
.
r
else
app
.
state
.
RELEVANCE_THRESHOLD
,
embeddings_function
=
embeddings_function
,
embeddings_function
=
embeddings_function
,
reranking_function
=
app
.
state
.
sentence_transformer_rf
,
reranking_function
=
app
.
state
.
sentence_transformer_rf
,
hybrid
=
form_data
.
hybrid
if
form_data
.
hybrid
else
app
.
state
.
HYBRID
,
)
)
except
Exception
as
e
:
except
Exception
as
e
:
log
.
exception
(
e
)
log
.
exception
(
e
)
...
...
backend/apps/rag/utils.py
View file @
9755cd5b
...
@@ -32,13 +32,13 @@ def query_embeddings_doc(
...
@@ -32,13 +32,13 @@ def query_embeddings_doc(
collection_name
:
str
,
collection_name
:
str
,
query
:
str
,
query
:
str
,
embeddings_function
,
embeddings_function
,
reranking_function
,
k
:
int
,
k
:
int
,
reranking_function
:
Optional
[
CrossEncoder
]
=
None
,
r
:
Optional
[
float
]
=
None
,
r
:
Optional
[
float
]
=
None
,
hybrid
:
Optional
[
bool
]
=
False
,
):
):
try
:
try
:
if
hybrid
:
if
reranking_function
:
# if you use docker use the model from the environment variable
# if you use docker use the model from the environment variable
collection
=
CHROMA_CLIENT
.
get_collection
(
name
=
collection_name
)
collection
=
CHROMA_CLIENT
.
get_collection
(
name
=
collection_name
)
...
@@ -142,6 +142,7 @@ def query_embeddings_collection(
...
@@ -142,6 +142,7 @@ def query_embeddings_collection(
r
:
float
,
r
:
float
,
embeddings_function
,
embeddings_function
,
reranking_function
,
reranking_function
,
hybrid
:
bool
,
):
):
results
=
[]
results
=
[]
...
@@ -155,6 +156,7 @@ def query_embeddings_collection(
...
@@ -155,6 +156,7 @@ def query_embeddings_collection(
r
=
r
,
r
=
r
,
embeddings_function
=
embeddings_function
,
embeddings_function
=
embeddings_function
,
reranking_function
=
reranking_function
,
reranking_function
=
reranking_function
,
hybrid
=
hybrid
,
)
)
results
.
append
(
result
)
results
.
append
(
result
)
except
:
except
:
...
@@ -211,6 +213,7 @@ def rag_messages(
...
@@ -211,6 +213,7 @@ def rag_messages(
template
,
template
,
k
,
k
,
r
,
r
,
hybrid
,
embedding_engine
,
embedding_engine
,
embedding_model
,
embedding_model
,
embedding_function
,
embedding_function
,
...
@@ -283,6 +286,7 @@ def rag_messages(
...
@@ -283,6 +286,7 @@ def rag_messages(
r
=
r
,
r
=
r
,
embeddings_function
=
embeddings_function
,
embeddings_function
=
embeddings_function
,
reranking_function
=
reranking_function
,
reranking_function
=
reranking_function
,
hybrid
=
hybrid
,
)
)
else
:
else
:
context
=
query_embeddings_doc
(
context
=
query_embeddings_doc
(
...
@@ -292,6 +296,7 @@ def rag_messages(
...
@@ -292,6 +296,7 @@ def rag_messages(
r
=
r
,
r
=
r
,
embeddings_function
=
embeddings_function
,
embeddings_function
=
embeddings_function
,
reranking_function
=
reranking_function
,
reranking_function
=
reranking_function
,
hybrid
=
hybrid
,
)
)
except
Exception
as
e
:
except
Exception
as
e
:
log
.
exception
(
e
)
log
.
exception
(
e
)
...
...
backend/config.py
View file @
9755cd5b
...
@@ -422,6 +422,7 @@ CHROMA_DATA_PATH = f"{DATA_DIR}/vector_db"
...
@@ -422,6 +422,7 @@ CHROMA_DATA_PATH = f"{DATA_DIR}/vector_db"
RAG_TOP_K
=
int
(
os
.
environ
.
get
(
"RAG_TOP_K"
,
"5"
))
RAG_TOP_K
=
int
(
os
.
environ
.
get
(
"RAG_TOP_K"
,
"5"
))
RAG_RELEVANCE_THRESHOLD
=
float
(
os
.
environ
.
get
(
"RAG_RELEVANCE_THRESHOLD"
,
"0.0"
))
RAG_RELEVANCE_THRESHOLD
=
float
(
os
.
environ
.
get
(
"RAG_RELEVANCE_THRESHOLD"
,
"0.0"
))
RAG_HYBRID
=
os
.
environ
.
get
(
"RAG_HYBRID"
,
""
).
lower
()
==
"true"
RAG_EMBEDDING_ENGINE
=
os
.
environ
.
get
(
"RAG_EMBEDDING_ENGINE"
,
""
)
RAG_EMBEDDING_ENGINE
=
os
.
environ
.
get
(
"RAG_EMBEDDING_ENGINE"
,
""
)
...
...
src/lib/components/documents/Settings/General.svelte
View file @
9755cd5b
...
@@ -43,7 +43,8 @@
...
@@ -43,7 +43,8 @@
let querySettings = {
let querySettings = {
template: '',
template: '',
r: 0.0,
r: 0.0,
k: 4
k: 4,
hybrid: false
};
};
const scanHandler = async () => {
const scanHandler = async () => {
...
@@ -174,6 +175,12 @@
...
@@ -174,6 +175,12 @@
}
}
};
};
const toggleHybridSearch = async () => {
querySettings.hybrid = !querySettings.hybrid;
querySettings = await updateQuerySettings(localStorage.token, querySettings);
};
onMount(async () => {
onMount(async () => {
const res = await getRAGConfig(localStorage.token);
const res = await getRAGConfig(localStorage.token);
...
@@ -202,6 +209,24 @@
...
@@ -202,6 +209,24 @@
<div>
<div>
<div class=" mb-2 text-sm font-medium">{$i18n.t('General Settings')}</div>
<div class=" mb-2 text-sm font-medium">{$i18n.t('General Settings')}</div>
<div class=" flex w-full justify-between">
<div class=" self-center text-xs font-medium">{$i18n.t('Hybrid Search')}</div>
<button
class="p-1 px-3 text-xs flex rounded transition"
on:click={() => {
toggleHybridSearch();
}}
type="button"
>
{#if querySettings.hybrid === true}
<span class="ml-2 self-center">{$i18n.t('On')}</span>
{:else}
<span class="ml-2 self-center">{$i18n.t('Off')}</span>
{/if}
</button>
</div>
<div class=" flex w-full justify-between">
<div class=" flex w-full justify-between">
<div class=" self-center text-xs font-medium">{$i18n.t('Embedding Model Engine')}</div>
<div class=" self-center text-xs font-medium">{$i18n.t('Embedding Model Engine')}</div>
<div class="flex items-center relative">
<div class="flex items-center relative">
...
@@ -386,78 +411,74 @@
...
@@ -386,78 +411,74 @@
<hr class=" dark:border-gray-700 my-3" />
<hr class=" dark:border-gray-700 my-3" />
<div class=" ">
{#if querySettings.hybrid === true}
<div class=" mb-2 text-sm font-medium">{$i18n.t('Update Reranking Model')}</div>
<div class=" ">
<div class=" mb-2 text-sm font-medium">{$i18n.t('Update Reranking Model')}</div>
<div class="flex w-full">
<div class="flex w-full">
<div class="flex-1 mr-2">
<div class="flex-1 mr-2">
<input
<input
class="w-full rounded-lg py-2 px-4 text-sm dark:text-gray-300 dark:bg-gray-850 outline-none"
class="w-full rounded-lg py-2 px-4 text-sm dark:text-gray-300 dark:bg-gray-850 outline-none"
placeholder={$i18n.t('Update reranking model (e.g. {{model}})', {
placeholder={$i18n.t('Update reranking model (e.g. {{model}})', {
model: rerankingModel.slice(-40)
model: rerankingModel.slice(-40)
})}
})}
bind:value={rerankingModel}
bind:value={rerankingModel}
/>
/>
</div>
</div>
<button
<button
class="px-2.5 bg-gray-100 hover:bg-gray-200 text-gray-800 dark:bg-gray-850 dark:hover:bg-gray-800 dark:text-gray-100 rounded-lg transition"
class="px-2.5 bg-gray-100 hover:bg-gray-200 text-gray-800 dark:bg-gray-850 dark:hover:bg-gray-800 dark:text-gray-100 rounded-lg transition"
on:click={() => {
on:click={() => {
rerankingModelUpdateHandler();
rerankingModelUpdateHandler();
}}
}}
disabled={updateRerankingModelLoading}
disabled={updateRerankingModelLoading}
>
>
{#if updateRerankingModelLoading}
{#if updateRerankingModelLoading}
<div class="self-center">
<div class="self-center">
<svg
class=" w-4 h-4"
viewBox="0 0 24 24"
fill="currentColor"
xmlns="http://www.w3.org/2000/svg"
><style>
.spinner_ajPY {
transform-origin: center;
animation: spinner_AtaB 0.75s infinite linear;
}
@keyframes spinner_AtaB {
100% {
transform: rotate(360deg);
}
}
</style><path
d="M12,1A11,11,0,1,0,23,12,11,11,0,0,0,12,1Zm0,19a8,8,0,1,1,8-8A8,8,0,0,1,12,20Z"
opacity=".25"
/><path
d="M10.14,1.16a11,11,0,0,0-9,8.92A1.59,1.59,0,0,0,2.46,12,1.52,1.52,0,0,0,4.11,10.7a8,8,0,0,1,6.66-6.61A1.42,1.42,0,0,0,12,2.69h0A1.57,1.57,0,0,0,10.14,1.16Z"
class="spinner_ajPY"
/></svg
>
</div>
{:else}
<svg
<svg
class=" w-4 h-4"
viewBox="0 0 24 24"
fill="currentColor"
xmlns="http://www.w3.org/2000/svg"
xmlns="http://www.w3.org/2000/svg"
><style>
viewBox="0 0 16 16"
.spinner_ajPY {
fill="currentColor"
transform-origin: center;
class="w-4 h-4"
animation: spinner_AtaB 0.75s infinite linear;
}
@keyframes spinner_AtaB {
100% {
transform: rotate(360deg);
}
}
</style><path
d="M12,1A11,11,0,1,0,23,12,11,11,0,0,0,12,1Zm0,19a8,8,0,1,1,8-8A8,8,0,0,1,12,20Z"
opacity=".25"
/><path
d="M10.14,1.16a11,11,0,0,0-9,8.92A1.59,1.59,0,0,0,2.46,12,1.52,1.52,0,0,0,4.11,10.7a8,8,0,0,1,6.66-6.61A1.42,1.42,0,0,0,12,2.69h0A1.57,1.57,0,0,0,10.14,1.16Z"
class="spinner_ajPY"
/></svg
>
>
</div>
<path
{:else}
d="M8.75 2.75a.75.75 0 0 0-1.5 0v5.69L5.03 6.22a.75.75 0 0 0-1.06 1.06l3.5 3.5a.75.75 0 0 0 1.06 0l3.5-3.5a.75.75 0 0 0-1.06-1.06L8.75 8.44V2.75Z"
<svg
/>
xmlns="http://www.w3.org/2000/svg"
<path
viewBox="0 0 16 16"
d="M3.5 9.75a.75.75 0 0 0-1.5 0v1.5A2.75 2.75 0 0 0 4.75 14h6.5A2.75 2.75 0 0 0 14 11.25v-1.5a.75.75 0 0 0-1.5 0v1.5c0 .69-.56 1.25-1.25 1.25h-6.5c-.69 0-1.25-.56-1.25-1.25v-1.5Z"
fill="currentColor"
/>
class="w-4 h-4"
</svg>
>
{/if}
<path
</button>
d="M8.75 2.75a.75.75 0 0 0-1.5 0v5.69L5.03 6.22a.75.75 0 0 0-1.06 1.06l3.5 3.5a.75.75 0 0 0 1.06 0l3.5-3.5a.75.75 0 0 0-1.06-1.06L8.75 8.44V2.75Z"
</div>
/>
<path
d="M3.5 9.75a.75.75 0 0 0-1.5 0v1.5A2.75 2.75 0 0 0 4.75 14h6.5A2.75 2.75 0 0 0 14 11.25v-1.5a.75.75 0 0 0-1.5 0v1.5c0 .69-.56 1.25-1.25 1.25h-6.5c-.69 0-1.25-.56-1.25-1.25v-1.5Z"
/>
</svg>
{/if}
</button>
</div>
</div>
</div>
<div class="mt-2 mb-1 text-xs text-gray-400 dark:text-gray-500">
{$i18n.t(
'Note: If you choose a reranking model, it will use that to score and rerank instead of the embedding model.'
)}
</div>
<hr class=" dark:border-gray-700 my-3" />
<hr class=" dark:border-gray-700 my-3" />
{/if}
<div class=" flex w-full justify-between">
<div class=" flex w-full justify-between">
<div class=" self-center text-xs font-medium">
<div class=" self-center text-xs font-medium">
...
@@ -583,25 +604,27 @@
...
@@ -583,25 +604,27 @@
</div>
</div>
</div>
</div>
<div class=" flex">
{#if querySettings.hybrid === true}
<div class=" flex w-full justify-between">
<div class=" flex">
<div class="self-center text-xs font-medium flex-1">
<div class=" flex w-full justify-between">
{$i18n.t('Relevance Threshold')}
<div class="self-center text-xs font-medium flex-1">
</div>
{$i18n.t('Relevance Threshold')}
</div>
<div class="self-center p-3">
<div class="self-center p-3">
<input
<input
class=" w-full rounded-lg py-1.5 px-4 text-sm dark:text-gray-300 dark:bg-gray-850 outline-none"
class=" w-full rounded-lg py-1.5 px-4 text-sm dark:text-gray-300 dark:bg-gray-850 outline-none"
type="number"
type="number"
step="0.01"
step="0.01"
placeholder={$i18n.t('Enter Relevance Threshold')}
placeholder={$i18n.t('Enter Relevance Threshold')}
bind:value={querySettings.r}
bind:value={querySettings.r}
autocomplete="off"
autocomplete="off"
min="0.0"
min="0.0"
/>
/>
</div>
</div>
</div>
</div>
</div>
</div>
{/if}
<div>
<div>
<div class=" mb-2.5 text-sm font-medium">{$i18n.t('RAG Template')}</div>
<div class=" mb-2.5 text-sm font-medium">{$i18n.t('RAG Template')}</div>
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment