Feature / collections (#96)

* Update schema defs for source -> metadata
* Migrate to use metadata part of schema, also add metadata to triples & vecs
* Add user/collection metadata to query
* Use user/collection in RAG
* Write and query working on triples
This commit is contained in:
cybermaggedon 2024-10-02 18:14:29 +01:00 committed by GitHub
parent 709221fa10
commit b0f4c58200
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
31 changed files with 459 additions and 251 deletions

View file

@ -6,7 +6,7 @@ Loads a text document into TrustGraph processing.
import pulsar
from pulsar.schema import JsonSchema
from trustgraph.schema import TextDocument, Source, text_ingest_queue
from trustgraph.schema import TextDocument, text_ingest_queue, Metadata
import base64
import hashlib
import argparse
@ -15,12 +15,17 @@ import time
from trustgraph.log_level import LogLevel
default_user = 'trustgraph'
default_collection = 'default'
class Loader:
def __init__(
self,
pulsar_host,
output_queue,
user,
collection,
log_level,
):
@ -35,6 +40,9 @@ class Loader:
chunking_enabled=True,
)
self.user = user
self.collection = collection
def load(self, files):
for file in files:
@ -50,10 +58,12 @@ class Loader:
id = hashlib.sha256(path.encode("utf-8")).hexdigest()[0:8]
r = TextDocument(
source=Source(
metadata=Metadata(
source=path,
title=path,
id=id,
user=self.user,
collection=self.collection,
),
text=data,
)
@ -90,6 +100,18 @@ def main():
help=f'Output queue (default: {default_output_queue})'
)
parser.add_argument(
'-u', '--user',
default=default_user,
help=f'User ID (default: {default_user})'
)
parser.add_argument(
'-c', '--collection',
default=default_collection,
help=f'Collection ID (default: {default_collection})'
)
parser.add_argument(
'-l', '--log-level',
type=LogLevel,
@ -112,6 +134,8 @@ def main():
p = Loader(
pulsar_host=args.pulsar_host,
output_queue=args.output_queue,
user=args.user,
collection=args.collection,
log_level=args.log_level,
)