Customization
Authentication
Check how we handle authentication in:
Observability
You can use sentry (opens in a new tab) for error reporting. You can set your own sentry config in config.yaml
config.yaml
sentry: YOUR_DSNCustom database
You need to implement the abstract class VectorDatabase found here (opens in a new tab) and then use it like:
from embedbase import get_app
from embedbase.database.supabase_db import Supabase
from embedbase.embedding.openai import OpenAI
from .my_custom_db import MyCustomDb
app = (
get_app()
.use_embedder(OpenAI("<your key>"))
.use_db(MyCustomDb())
).run()Custom middleware
Example production middlewares:
- Playground Embedbase instance is open source (opens in a new tab)
You can a custom middleware like this:
from embedbase import get_app
from embedbase.database.postgres_db import Postgres
from embedbase.embedding.openai import OpenAI
async def process_time(request, call_next):
start_time = time.time()
response = await call_next(request)
process_time = time.time() - start_time
response.headers["X-Process-Time"] = str(process_time)
return response
app = (
get_app()
.use_middleware(process_time)
.use_embedder(OpenAI("<your key>"))
.use_db(Postgres())
).run()
curl -X POST -H "Content-Type: application/json" -d '{"query": "Bob"}' http://localhost:8080/v1/people/search{
"query": "Bob",
"similarities": [
{
"score": 0.828773,
"id": "ABCU75FEBE",
"data": "Elon is sipping a tea on Mars",
}
],
"headers": {
"x-process-time": "0.0001239776611328125"
}
}Custom Embedder
You need to implement the abstract class Embedder found here (opens in a new tab) and then use it like:
from embedbase import get_app
from embedbase.database.postgres_db import Postgres
from .my_custom_embedder import MyCustomEmbedder
app = (
get_app()
.use_db(Postgres())
.use_embedder(MyCustomEmbedder())
).run()Check this example (opens in a new tab).