apunkt/SurfSense
1
0
Fork 0
mirror of https://github.com/MODSetter/SurfSense.git synced 2026-06-10 20:35:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

Merge pull request #408 from MODSetter/dev

Browse source 
chore: updated docs & refactored sst_service.py
This commit is contained in:
Rohan Verma 2025-10-15 14:44:03 -07:00 • committed by GitHub
commit fa39176b82
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
10 changed files with 195 additions and 173 deletions
Download patch file Download diff file Expand all files Collapse all files

31
surfsense_backend/app/routes/documents_routes.py
View file

@ -785,20 +785,25 @@ async def process_file_in_background(
)
# Determine STT service type
stt_service_type = "local" if app_config.STT_SERVICE and app_config.STT_SERVICE.startswith("local/") else "external"
stt_service_type = (
"local"
if app_config.STT_SERVICE
and app_config.STT_SERVICE.startswith("local/")
else "external"
)
# Check if using local STT service
if stt_service_type == "local":
# Use local Faster-Whisper for transcription
from app.services.stt_service import stt_service
try:
result = stt_service.transcribe_file(file_path)
transcribed_text = result.get("text", "")
if not transcribed_text:
raise ValueError("Transcription returned empty text")
# Add metadata about the transcription
transcribed_text = (
f"# Transcription of {filename}\n\n{transcribed_text}"
@ -806,9 +811,9 @@ async def process_file_in_background(
except Exception as e:
raise HTTPException(
status_code=422,
detail=f"Failed to transcribe audio file {filename}: {str(e)}"
detail=f"Failed to transcribe audio file {filename}: {e!s}",
) from e
await task_logger.log_task_progress(
log_entry,
f"Local STT transcription completed: {filename}",
@ -828,13 +833,17 @@ async def process_file_in_background(
"api_key": app_config.STT_SERVICE_API_KEY,
}
if app_config.STT_SERVICE_API_BASE:
transcription_kwargs["api_base"] = app_config.STT_SERVICE_API_BASE
transcription_response = await atranscription(**transcription_kwargs)
transcription_kwargs["api_base"] = (
app_config.STT_SERVICE_API_BASE
)
transcription_response = await atranscription(
**transcription_kwargs
)
# Extract the transcribed text
transcribed_text = transcription_response.get("text", "")
if not transcribed_text:
raise ValueError("Transcription returned empty text")

44
surfsense_backend/app/services/stt_service.py
View file

@ -3,15 +3,15 @@
import os
import tempfile
from pathlib import Path
from typing import Optional
from faster_whisper import WhisperModel
from app.config import config
class STTService:
"""Local Speech-to-Text service using Faster-Whisper."""
def __init__(self):
"""Initialize STT service with model from STT_SERVICE config."""
# Parse model from STT_SERVICE (e.g., "local/base" or "local/tiny")
@ -20,8 +20,8 @@ class STTService:
self.model_size = stt_service.split("/", 1)[1]
else:
self.model_size = "base" # fallback
self._model: Optional[WhisperModel] = None
self._model: WhisperModel | None = None
def _get_model(self) -> WhisperModel:
"""Lazy load the Whisper model."""
if self._model is None:
@ -33,49 +33,53 @@ class STTService:
num_workers=1, # Single worker for stability
)
return self._model
def transcribe_file(self, audio_path: str, language: Optional[str] = None) -> dict:
def transcribe_file(self, audio_path: str, language: str | None = None) -> dict:
"""Transcribe audio file to text.
Args:
audio_path: Path to audio file
language: Optional language code (e.g., "en", "es")
Returns:
Dict with transcription text and metadata
"""
model = self._get_model()
# Transcribe with optimized settings
segments, info = model.transcribe(
audio_path,
language=language,
beam_size=1, # Faster inference
best_of=1, # Single pass
best_of=1, # Single pass
temperature=0, # Deterministic output
vad_filter=True, # Voice activity detection
vad_parameters=dict(min_silence_duration_ms=500),
vad_parameters={"min_silence_duration_ms": 500},
)
# Combine all segments
text = " ".join(segment.text.strip() for segment in segments)
return {
"text": text,
"language": info.language,
"language_probability": info.language_probability,
"duration": info.duration,
}
def transcribe_bytes(self, audio_bytes: bytes, filename: str = "audio.wav",
language: Optional[str] = None) -> dict:
def transcribe_bytes(
self,
audio_bytes: bytes,
filename: str = "audio.wav",
language: str | None = None,
) -> dict:
"""Transcribe audio from bytes.
Args:
audio_bytes: Audio file bytes
filename: Original filename for format detection
language: Optional language code
Returns:
Dict with transcription text and metadata
"""
@ -84,7 +88,7 @@ class STTService:
with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as tmp_file:
tmp_file.write(audio_bytes)
tmp_path = tmp_file.name
try:
return self.transcribe_file(tmp_path, language)
finally:
@ -93,4 +97,4 @@ class STTService:
# Global STT service instance
stt_service = STTService()
stt_service = STTService()

2
surfsense_web/app/dashboard/[search_space_id]/documents/(manage)/page.tsx
View file

@ -36,7 +36,7 @@ export default function DocumentsTable() {
created_at: true,
});
const [pageIndex, setPageIndex] = useState(0);
const [pageSize, setPageSize] = useState(10);
const [pageSize, setPageSize] = useState(50);
const [sortKey, setSortKey] = useState<SortKey>("title");
const [sortDesc, setSortDesc] = useState(false);
const [selectedIds, setSelectedIds] = useState<Set<number>>(new Set());

8
surfsense_web/components/chat/ChatInputGroup.tsx
View file

@ -334,9 +334,13 @@ ResearchModeSelector.displayName = "ResearchModeSelector";
const LLMSelector = React.memo(() => {
const { search_space_id } = useParams();
const searchSpaceId = Number(search_space_id);
const { llmConfigs, loading: llmLoading, error } = useLLMConfigs(searchSpaceId);
const { preferences, updatePreferences, loading: preferencesLoading } = useLLMPreferences(searchSpaceId);
const {
preferences,
updatePreferences,
loading: preferencesLoading,
} = useLLMPreferences(searchSpaceId);
const isLoading = llmLoading || preferencesLoading;

235
surfsense_web/components/inference-params-editor.tsx
View file

@ -1,138 +1,153 @@
"use client";
import { useState } from "react";
import { Plus, Trash2 } from "lucide-react";
import { useState } from "react";
import { Button } from "@/components/ui/button";
import { Input } from "@/components/ui/input";
import { Label } from "@/components/ui/label";
import {
Select,
SelectContent,
SelectItem,
SelectTrigger,
SelectValue,
Select,
SelectContent,
SelectItem,
SelectTrigger,
SelectValue,
} from "@/components/ui/select";
interface InferenceParamsEditorProps {
params: Record<string, number | string>;
setParams: (newParams: Record<string, number | string>) => void;
params: Record<string, number | string>;
setParams: (newParams: Record<string, number | string>) => void;
}
const PARAM_KEYS = ["temperature", "max_tokens", "top_k", "top_p"] as const;
export default function InferenceParamsEditor({
params,
setParams,
}: InferenceParamsEditorProps) {
const [selectedKey, setSelectedKey] = useState<string>("");
const [value, setValue] = useState<string>("");
export default function InferenceParamsEditor({ params, setParams }: InferenceParamsEditorProps) {
const [selectedKey, setSelectedKey] = useState<string>("");
const [value, setValue] = useState<string>("");
const handleAdd = () => {
if (!selectedKey || value === "") return;
const handleAdd = () => {
if (!selectedKey || value === "") return;
if (params[selectedKey]) {
alert(`${selectedKey} already exists`);
return;
}
if (params[selectedKey]) {
alert(`${selectedKey} already exists`);
return;
}
const numericValue = Number(value);
const numericValue = Number(value);
if ((selectedKey === "temperature" || selectedKey === "top_p") && (isNaN(numericValue) || numericValue < 0 || numericValue > 1)) {
alert("Value must be a number between 0 and 1");
return;
}
if (
(selectedKey === "temperature" || selectedKey === "top_p") &&
(isNaN(numericValue) || numericValue < 0 || numericValue > 1)
) {
alert("Value must be a number between 0 and 1");
return;
}
if ((selectedKey === "max_tokens" || selectedKey === "top_k") && (!Number.isInteger(numericValue) || numericValue < 0)) {
alert("Value must be a non-negative integer");
return;
}
if (
(selectedKey === "max_tokens" || selectedKey === "top_k") &&
(!Number.isInteger(numericValue) || numericValue < 0)
) {
alert("Value must be a non-negative integer");
return;
}
setParams({
...params,
[selectedKey]: isNaN(numericValue) ? value : numericValue,
});
setParams({
...params,
[selectedKey]: isNaN(numericValue) ? value : numericValue,
});
setSelectedKey("");
setValue("");
};
setSelectedKey("");
setValue("");
};
const handleDelete = (key: string) => {
const newParams = { ...params };
delete newParams[key];
setParams(newParams);
};
const handleDelete = (key: string) => {
const newParams = { ...params };
delete newParams[key];
setParams(newParams);
};
return (
<div className="space-y-6 p-2 sm:p-0">
<div className="grid grid-cols-1 gap-4 md:grid-cols-[1fr_1fr_auto] md:gap-3 items-end">
<div className="flex flex-col space-y-1">
<Label htmlFor="param-key" className="text-sm font-medium">Parameter Key</Label>
<Select value={selectedKey} onValueChange={setSelectedKey}>
<SelectTrigger id="param-key" className="w-full">
<SelectValue placeholder="Select parameter" />
</SelectTrigger>
<SelectContent>
{PARAM_KEYS.map((key) => (
<SelectItem key={key} value={key}>{key}</SelectItem>
))}
</SelectContent>
</Select>
</div>
return (
<div className="space-y-6 p-2 sm:p-0">
<div className="grid grid-cols-1 gap-4 md:grid-cols-[1fr_1fr_auto] md:gap-3 items-end">
<div className="flex flex-col space-y-1">
<Label htmlFor="param-key" className="text-sm font-medium">
Parameter Key
</Label>
<Select value={selectedKey} onValueChange={setSelectedKey}>
<SelectTrigger id="param-key" className="w-full">
<SelectValue placeholder="Select parameter" />
</SelectTrigger>
<SelectContent>
{PARAM_KEYS.map((key) => (
<SelectItem key={key} value={key}>
{key}
</SelectItem>
))}
</SelectContent>
</Select>
</div>
<div className="flex flex-col space-y-1">
<Label htmlFor="param-value" className="text-sm font-medium">Value</Label>
<Input
id="param-value"
placeholder="Enter value (e.g., 0.7 or 512)"
value={value}
onChange={(e) => setValue(e.target.value)}
className="w-full"
/>
</div>
<div className="flex flex-col space-y-1">
<Label htmlFor="param-value" className="text-sm font-medium">
Value
</Label>
<Input
id="param-value"
placeholder="Enter value (e.g., 0.7 or 512)"
value={value}
onChange={(e) => setValue(e.target.value)}
className="w-full"
/>
</div>
<Button
className="w-full md:w-auto h-10 mt-0"
onClick={handleAdd}
disabled={!selectedKey || value === ""}
>
<Plus className="w-4 h-4 mr-2" /> Add Parameter
</Button>
</div>
<Button
className="w-full md:w-auto h-10 mt-0"
onClick={handleAdd}
disabled={!selectedKey || value === ""}
>
<Plus className="w-4 h-4 mr-2" /> Add Parameter
</Button>
</div>
<hr className="my-4" />
<hr className="my-4" />
{Object.keys(params).length > 0 && (
<div className="border rounded-lg shadow-sm overflow-x-auto">
<table className="min-w-full text-left text-sm divide-y divide-gray-200">
<thead className="bg-black dark:bg-black">
<tr>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300">Key</th>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300">Value</th>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300 sr-only md:not-sr-only">Actions</th>
</tr>
</thead>
<tbody className="divide-y divide-gray-200 bg-black dark:bg-black">
{Object.entries(params).map(([key, val]) => (
<tr key={key} className="hover:bg-gray-50 dark:hover:bg-gray-800 transition-colors">
<td className="px-4 py-3 font-medium text-gray-900 dark:text-white">{key}</td>
<td className="px-4 py-3 text-gray-700 dark:text-gray-300">{val.toString()}</td>
<td className="px-4 py-3">
<Button
variant="ghost"
size="icon"
className="text-red-600 hover:bg-red-50 dark:hover:bg-red-900/20 hover:text-red-700 dark:text-red-500"
onClick={() => handleDelete(key)}
aria-label={`Delete parameter ${key}`}
>
<Trash2 className="w-4 h-4" />
</Button>
</td>
</tr>
))}
</tbody>
</table>
</div>
)}
</div>
);
{Object.keys(params).length > 0 && (
<div className="border rounded-lg shadow-sm overflow-x-auto">
<table className="min-w-full text-left text-sm divide-y divide-gray-200">
<thead className="bg-black dark:bg-black">
<tr>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300">
Key
</th>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300">
Value
</th>
<th className="px-4 py-3 text-xs font-semibold uppercase tracking-wider text-gray-600 dark:text-gray-300 sr-only md:not-sr-only">
Actions
</th>
</tr>
</thead>
<tbody className="divide-y divide-gray-200 bg-black dark:bg-black">
{Object.entries(params).map(([key, val]) => (
<tr key={key} className="hover:bg-gray-50 dark:hover:bg-gray-800 transition-colors">
<td className="px-4 py-3 font-medium text-gray-900 dark:text-white">{key}</td>
<td className="px-4 py-3 text-gray-700 dark:text-gray-300">{val.toString()}</td>
<td className="px-4 py-3">
<Button
variant="ghost"
size="icon"
className="text-red-600 hover:bg-red-50 dark:hover:bg-red-900/20 hover:text-red-700 dark:text-red-500"
onClick={() => handleDelete(key)}
aria-label={`Delete parameter ${key}`}
>
<Trash2 className="w-4 h-4" />
</Button>
</td>
</tr>
))}
</tbody>
</table>
</div>
)}
</div>
);
}

3
surfsense_web/components/onboard/add-provider-step.tsx
View file

@ -17,8 +17,8 @@ import {
SelectTrigger,
SelectValue,
} from "@/components/ui/select";
import { LLM_PROVIDERS } from "@/contracts/enums/llm-providers";
import { LANGUAGES } from "@/contracts/enums/languages";
import { LLM_PROVIDERS } from "@/contracts/enums/llm-providers";
import { type CreateLLMConfig, useLLMConfigs } from "@/hooks/use-llm-configs";
import InferenceParamsEditor from "../inference-params-editor";
@ -223,7 +223,6 @@ export function AddProviderStep({
</SelectContent>
</Select>
</div>
</div>
{formData.provider === "CUSTOM" && (

6
surfsense_web/content/docs/docker-installation.mdx
View file

@ -85,10 +85,10 @@ Before you begin, ensure you have:
| RERANKERS_MODEL_NAME | Name of the reranker model (e.g., `ms-marco-MiniLM-L-12-v2`) |
| RERANKERS_MODEL_TYPE | Type of reranker model (e.g., `flashrank`) |
| TTS_SERVICE | Text-to-Speech API provider for Podcasts (e.g., `local/kokoro`, `openai/tts-1`). See [supported providers](https://docs.litellm.ai/docs/text_to_speech#supported-providers) |
| TTS_SERVICE_API_KEY | API key for the Text-to-Speech service |
| TTS_SERVICE_API_KEY | (Optional if local) API key for the Text-to-Speech service |
| TTS_SERVICE_API_BASE | (Optional) Custom API base URL for the Text-to-Speech service |
| STT_SERVICE | Speech-to-Text API provider for Podcasts (e.g., `openai/whisper-1`). See [supported providers](https://docs.litellm.ai/docs/audio_transcription#supported-providers) |
| STT_SERVICE_API_KEY | API key for the Speech-to-Text service |
| STT_SERVICE | Speech-to-Text API provider for Audio Files (e.g., `local/base`, `openai/whisper-1`). See [supported providers](https://docs.litellm.ai/docs/audio_transcription#supported-providers) |
| STT_SERVICE_API_KEY | (Optional if local) API key for the Speech-to-Text service |
| STT_SERVICE_API_BASE | (Optional) Custom API base URL for the Speech-to-Text service |
| FIRECRAWL_API_KEY | API key for Firecrawl service for web crawling |
| ETL_SERVICE | Document parsing service: `UNSTRUCTURED` (supports 34+ formats), `LLAMACLOUD` (supports 50+ formats including legacy document types), or `DOCLING` (local processing, supports PDF, Office docs, images, HTML, CSV) |

11
surfsense_web/content/docs/manual-installation.mdx
View file

@ -62,12 +62,11 @@ Edit the `.env` file and set the following variables:
| RERANKERS_MODEL_NAME | Name of the reranker model (e.g., `ms-marco-MiniLM-L-12-v2`) |
| RERANKERS_MODEL_TYPE | Type of reranker model (e.g., `flashrank`) |
| TTS_SERVICE | Text-to-Speech API provider for Podcasts (e.g., `local/kokoro`, `openai/tts-1`). See [supported providers](https://docs.litellm.ai/docs/text_to_speech#supported-providers) |
| TTS_SERVICE_API_KEY | API key for the Text-to-Speech service |
| TTS_SERVICE_API_BASE | (Optional) Custom API base URL for the Text-to-Speech service |
| STT_SERVICE | Speech-to-Text API provider for Podcasts (e.g., `openai/whisper-1`). See [supported providers](https://docs.litellm.ai/docs/audio_transcription#supported-providers) |
| STT_SERVICE_API_KEY | API key for the Speech-to-Text service |
| STT_SERVICE_API_BASE | (Optional) Custom API base URL for the Speech-to-Text service |
| FIRECRAWL_API_KEY | API key for Firecrawl service for web crawling |
| TTS_SERVICE_API_KEY | (Optional if local) API key for the Text-to-Speech service |
| TTS_SERVICE_API_BASE | (Optional) Custom API base URL for the Text-to-Speech service |
| STT_SERVICE | Speech-to-Text API provider for Audio Files (e.g., `local/base`, `openai/whisper-1`). See [supported providers](https://docs.litellm.ai/docs/audio_transcription#supported-providers) |
| STT_SERVICE_API_KEY | (Optional if local) API key for the Speech-to-Text service |
| STT_SERVICE_API_BASE | (Optional) Custom API base URL for the Speech-to-Text service |
| ETL_SERVICE | Document parsing service: `UNSTRUCTURED` (supports 34+ formats), `LLAMACLOUD` (supports 50+ formats including legacy document types), or `DOCLING` (local processing, supports PDF, Office docs, images, HTML, CSV) |
| UNSTRUCTURED_API_KEY | API key for Unstructured.io service for document parsing (required if ETL_SERVICE=UNSTRUCTURED) |
| LLAMA_CLOUD_API_KEY | API key for LlamaCloud service for document parsing (required if ETL_SERVICE=LLAMACLOUD) |

16
surfsense_web/drizzle.config.ts
View file

@ -1,11 +1,11 @@
import 'dotenv/config';
import { defineConfig } from 'drizzle-kit';
import "dotenv/config";
import { defineConfig } from "drizzle-kit";
export default defineConfig({
out: './drizzle',
schema: './app/db/schema.ts',
dialect: 'postgresql',
dbCredentials: {
url: process.env.DATABASE_URL!,
},
out: "./drizzle",
schema: "./app/db/schema.ts",
dialect: "postgresql",
dbCredentials: {
url: process.env.DATABASE_URL!,
},
});

12
surfsense_web/hooks/use-connector-edit-page.ts
View file

@ -337,12 +337,7 @@ export function useConnectorEditPage(connectorId: number, searchSpaceId: string)
const originalSafesearch = originalConfig.SEARXNG_SAFESEARCH;
if (safesearchRaw) {
const parsed = Number(safesearchRaw);
if (
Number.isNaN(parsed) ||
!Number.isInteger(parsed) ||
parsed < 0 ||
parsed > 2
) {
if (Number.isNaN(parsed) || !Number.isInteger(parsed) || parsed < 0 || parsed > 2) {
toast.error("SearxNG SafeSearch must be 0, 1, or 2.");
setIsSaving(false);
return;
@ -521,10 +516,7 @@ export function useConnectorEditPage(connectorId: number, searchSpaceId: string)
"SEARXNG_CATEGORIES",
normalizeListInput(newlySavedConfig.SEARXNG_CATEGORIES).join(", ")
);
editForm.setValue(
"SEARXNG_LANGUAGE",
newlySavedConfig.SEARXNG_LANGUAGE || ""
);
editForm.setValue("SEARXNG_LANGUAGE", newlySavedConfig.SEARXNG_LANGUAGE || "");
editForm.setValue(
"SEARXNG_SAFESEARCH",
newlySavedConfig.SEARXNG_SAFESEARCH === null ||