mirror of
https://github.com/trustgraph-ai/trustgraph.git
synced 2026-04-25 08:26:21 +02:00
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:
parent
709221fa10
commit
b0f4c58200
31 changed files with 459 additions and 251 deletions
|
|
@ -9,12 +9,17 @@ import os
|
|||
from trustgraph.clients.triples_query_client import TriplesQueryClient
|
||||
|
||||
default_pulsar_host = os.getenv("PULSAR_HOST", 'pulsar://localhost:6650')
|
||||
default_user = 'trustgraph'
|
||||
default_collection = 'default'
|
||||
|
||||
def show_graph(pulsar):
|
||||
def show_graph(pulsar, user, collection):
|
||||
|
||||
tq = TriplesQueryClient(pulsar_host=pulsar)
|
||||
|
||||
rows = tq.request(None, None, None, limit=10_000_000)
|
||||
rows = tq.request(
|
||||
user=user, collection=collection,
|
||||
s=None, p=None, o=None, limit=10_000_000
|
||||
)
|
||||
|
||||
for row in rows:
|
||||
print(row.s.value, row.p.value, row.o.value)
|
||||
|
|
@ -32,11 +37,26 @@ def main():
|
|||
help=f'Pulsar host (default: {default_pulsar_host})',
|
||||
)
|
||||
|
||||
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})'
|
||||
)
|
||||
|
||||
args = parser.parse_args()
|
||||
|
||||
try:
|
||||
|
||||
show_graph(args.pulsar_host)
|
||||
show_graph(
|
||||
pulsar=args.pulsar_host, user=args.user,
|
||||
collection=args.collection,
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
|
||||
|
|
|
|||
|
|
@ -6,7 +6,7 @@ Loads a PDF document into TrustGraph processing.
|
|||
|
||||
import pulsar
|
||||
from pulsar.schema import JsonSchema
|
||||
from trustgraph.schema import Document, Source, document_ingest_queue
|
||||
from trustgraph.schema import Document, document_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 = Document(
|
||||
source=Source(
|
||||
metadata=Metadata(
|
||||
source=path,
|
||||
title=path,
|
||||
id=id,
|
||||
user=self.user,
|
||||
collection=self.collection,
|
||||
),
|
||||
data=base64.b64encode(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,
|
||||
)
|
||||
|
||||
|
|
|
|||
|
|
@ -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,
|
||||
)
|
||||
|
||||
|
|
|
|||
|
|
@ -9,17 +9,19 @@ import os
|
|||
from trustgraph.clients.document_rag_client import DocumentRagClient
|
||||
|
||||
default_pulsar_host = os.getenv("PULSAR_HOST", 'pulsar://localhost:6650')
|
||||
default_user = 'trustgraph'
|
||||
default_collection = 'default'
|
||||
|
||||
def query(pulsar, query):
|
||||
def query(pulsar_host, query, user, collection):
|
||||
|
||||
rag = DocumentRagClient(pulsar_host=pulsar)
|
||||
resp = rag.request(query)
|
||||
resp = rag.request(user=user, collection=collection, query=query)
|
||||
print(resp)
|
||||
|
||||
def main():
|
||||
|
||||
parser = argparse.ArgumentParser(
|
||||
prog='graph-show',
|
||||
prog='tg-query-document-rag',
|
||||
description=__doc__,
|
||||
)
|
||||
|
||||
|
|
@ -35,11 +37,28 @@ def main():
|
|||
help=f'Query to execute',
|
||||
)
|
||||
|
||||
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})'
|
||||
)
|
||||
|
||||
args = parser.parse_args()
|
||||
|
||||
try:
|
||||
|
||||
query(args.pulsar_host, args.query)
|
||||
query(
|
||||
pulsar_host=args.pulsar_host,
|
||||
query=args.query,
|
||||
user=args.user,
|
||||
collection=args.collection,
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
|
||||
|
|
|
|||
|
|
@ -9,17 +9,19 @@ import os
|
|||
from trustgraph.clients.graph_rag_client import GraphRagClient
|
||||
|
||||
default_pulsar_host = os.getenv("PULSAR_HOST", 'pulsar://localhost:6650')
|
||||
default_user = 'trustgraph'
|
||||
default_collection = 'default'
|
||||
|
||||
def query(pulsar, query):
|
||||
def query(pulsar_host, query, user, collection):
|
||||
|
||||
rag = GraphRagClient(pulsar_host=pulsar)
|
||||
resp = rag.request(query)
|
||||
rag = GraphRagClient(pulsar_host=pulsar_host)
|
||||
resp = rag.request(user=user, collection=collection, query=query)
|
||||
print(resp)
|
||||
|
||||
def main():
|
||||
|
||||
parser = argparse.ArgumentParser(
|
||||
prog='graph-show',
|
||||
prog='tg-graph-query-rag',
|
||||
description=__doc__,
|
||||
)
|
||||
|
||||
|
|
@ -35,11 +37,28 @@ def main():
|
|||
help=f'Query to execute',
|
||||
)
|
||||
|
||||
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})'
|
||||
)
|
||||
|
||||
args = parser.parse_args()
|
||||
|
||||
try:
|
||||
|
||||
query(args.pulsar_host, args.query)
|
||||
query(
|
||||
pulsar_host=args.pulsar_host,
|
||||
query=args.query,
|
||||
user=args.user,
|
||||
collection=args.collection,
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue