trustgraph-cli 2.9.11__tar.gz → 2.10.1__tar.gz

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: trustgraph-cli
-Version: 2.9.11
+Version: 2.10.1
 Summary: TrustGraph provides a means to run a pipeline of flexible AI processing components in a flexible means to achieve a processing pipeline.
 Author-email: "trustgraph.ai" <security@trustgraph.ai>
 Project-URL: Homepage, https://github.com/trustgraph-ai/trustgraph
@@ -8,7 +8,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Operating System :: OS Independent
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
-Requires-Dist: trustgraph-base<2.10,>=2.9
+Requires-Dist: trustgraph-base<2.11,>=2.10
 Requires-Dist: requests
 Requires-Dist: pulsar-client
 Requires-Dist: aiohttp

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/pyproject.toml

@@ -10,7 +10,7 @@ description = "TrustGraph provides a means to run a pipeline of flexible AI proc
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
-    "trustgraph-base>=2.9,<2.10",
+    "trustgraph-base>=2.10,<2.11",
     "requests",
     "pulsar-client",
     "aiohttp",

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/invoke_document_embeddings.py

@@ -11,7 +11,10 @@ default_url = os.getenv("TRUSTGRAPH_URL", 'http://localhost:8888/')
 default_token = os.getenv("TRUSTGRAPH_TOKEN", None)
 default_workspace = os.getenv("TRUSTGRAPH_WORKSPACE", "default")
 
-def query(url, flow_id, query_text, collection, limit, token=None, workspace="default"):
+def query(
+    url, flow_id, query_text, collection, limit,
+    attributes=None, token=None, workspace="default",
+):
 
     # Create API client
     api = Api(url=url, token=token, workspace=workspace)
@@ -23,7 +26,8 @@ def query(url, flow_id, query_text, collection, limit, token=None, workspace="de
         result = flow.document_embeddings_query(
             text=query_text,
             collection=collection,
-            limit=limit
+            limit=limit,
+            attributes=attributes,
         )
 
         chunks = result.get("chunks", [])
@@ -83,6 +87,13 @@ def main():
         help='Maximum number of results (default: 10)',
     )
 
+    parser.add_argument(
+        '-a', '--attribute',
+        action='append',
+        default=[],
+        help='Filter attribute as key=value (repeatable)',
+    )
+
     parser.add_argument(
         'query',
         nargs=1,
@@ -91,6 +102,11 @@ def main():
 
     args = parser.parse_args()
 
+    attributes = {}
+    for attr in args.attribute:
+        k, _, v = attr.partition('=')
+        attributes[k] = v
+
     try:
 
         query(
@@ -99,6 +115,7 @@ def main():
             query_text=args.query[0],
             collection=args.collection,
             limit=args.limit,
+            attributes=attributes or None,
             token=args.token,
             workspace=args.workspace,
         )

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/invoke_graph_embeddings.py

@@ -4,6 +4,7 @@ Returns a list of matching graph entities.
 """
 
 import argparse
+import json
 import os
 from trustgraph.api import Api
 
@@ -11,7 +12,11 @@ default_url = os.getenv("TRUSTGRAPH_URL", 'http://localhost:8888/')
 default_token = os.getenv("TRUSTGRAPH_TOKEN", None)
 default_workspace = os.getenv("TRUSTGRAPH_WORKSPACE", "default")
 
-def query(url, flow_id, query_text, collection, limit, token=None, workspace="default"):
+def query(
+    url, flow_id, query_text, collection, limit,
+    rdf_type=None, attributes=None,
+    token=None, workspace="default",
+):
 
     # Create API client
     api = Api(url=url, token=token, workspace=workspace)
@@ -23,7 +28,9 @@ def query(url, flow_id, query_text, collection, limit, token=None, workspace="de
         result = flow.graph_embeddings_query(
             text=query_text,
             collection=collection,
-            limit=limit
+            limit=limit,
+            rdf_type=rdf_type,
+            attributes=attributes,
         )
 
         entities = result.get("entities", [])
@@ -93,6 +100,19 @@ def main():
         help='Maximum number of results (default: 10)',
     )
 
+    parser.add_argument(
+        '--rdf-type',
+        default=None,
+        help='RDF type IRI to filter results (optional)',
+    )
+
+    parser.add_argument(
+        '-a', '--attribute',
+        action='append',
+        default=[],
+        help='Filter attribute as key=value (repeatable)',
+    )
+
     parser.add_argument(
         'query',
         nargs=1,
@@ -101,6 +121,11 @@ def main():
 
     args = parser.parse_args()
 
+    attributes = {}
+    for attr in args.attribute:
+        k, _, v = attr.partition('=')
+        attributes[k] = v
+
     try:
 
         query(
@@ -109,6 +134,8 @@ def main():
             query_text=args.query[0],
             collection=args.collection,
             limit=args.limit,
+            rdf_type=args.rdf_type,
+            attributes=attributes or None,
             token=args.token,
             workspace=args.workspace,
         )

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/invoke_row_embeddings.py

@@ -11,7 +11,10 @@ default_url = os.getenv("TRUSTGRAPH_URL", 'http://localhost:8888/')
 default_token = os.getenv("TRUSTGRAPH_TOKEN", None)
 default_workspace = os.getenv("TRUSTGRAPH_WORKSPACE", "default")
 
-def query(url, flow_id, query_text, schema_name, collection, index_name, limit, token=None, workspace="default"):
+def query(
+    url, flow_id, query_text, schema_name, collection, index_name, limit,
+    attributes=None, token=None, workspace="default",
+):
 
     # Create API client
     api = Api(url=url, token=token, workspace=workspace)
@@ -25,6 +28,7 @@ def query(url, flow_id, query_text, schema_name, collection, index_name, limit,
             schema_name=schema_name,
             collection=collection,
             index_name=index_name,
+            attributes=attributes,
             limit=limit
         )
 
@@ -96,6 +100,13 @@ def main():
         help='Maximum number of results (default: 10)',
     )
 
+    parser.add_argument(
+        '-a', '--attribute',
+        action='append',
+        default=[],
+        help='Filter attribute as key=value (repeatable)',
+    )
+
     parser.add_argument(
         'query',
         nargs=1,
@@ -104,6 +115,11 @@ def main():
 
     args = parser.parse_args()
 
+    attributes = {}
+    for attr in args.attribute:
+        k, _, v = attr.partition('=')
+        attributes[k] = v
+
     try:
 
         query(
@@ -114,6 +130,7 @@ def main():
             collection=args.collection,
             index_name=args.index_name,
             limit=args.limit,
+            attributes=attributes or None,
             token=args.token,
             workspace=args.workspace,
         )

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/load_knowledge.py

@@ -9,6 +9,7 @@ import rdflib
 from typing import Iterator, Tuple
 
 from trustgraph.api import Api, Triple
+from trustgraph.api.types import Uri, Literal
 from trustgraph.log_level import LogLevel
 
 default_url = os.getenv("TRUSTGRAPH_URL", 'http://localhost:8888/')
@@ -26,6 +27,7 @@ class KnowledgeLoader:
             document_id,
             url=default_url,
             token=None, workspace="default",
+            graph="",
     ):
         self.files = files
         self.flow = flow
@@ -34,6 +36,7 @@ class KnowledgeLoader:
         self.url = url
         self.token = token
         self.workspace = workspace
+        self.graph = graph
 
     def load_triples_from_file(self, file) -> Iterator[Triple]:
         """Generator that yields Triple objects from a Turtle file"""
@@ -42,23 +45,19 @@ class KnowledgeLoader:
         g.parse(file, format="turtle")
 
         for e in g:
-            s_value = str(e[0])
-            p_value = str(e[1])
-            o_value = str(e[2])
-
-            o_datatype = ""
-            o_language = ""
+            s_value = Uri(str(e[0]))
+            p_value = Uri(str(e[1]))
 
             if isinstance(e[2], rdflib.term.Literal):
-                if e[2].language:
-                    o_language = str(e[2].language)
-                if e[2].datatype:
-                    o_datatype = str(e[2].datatype)
-
-            yield Triple(
-                s=s_value, p=p_value, o=o_value,
-                o_datatype=o_datatype, o_language=o_language,
-            )
+                o_value = Literal(
+                    str(e[2]),
+                    datatype=str(e[2].datatype) if e[2].datatype else None,
+                    language=str(e[2].language) if e[2].language else None,
+                )
+            else:
+                o_value = Uri(str(e[2]))
+
+            yield Triple(s=s_value, p=p_value, o=o_value, g=self.graph)
 
     def load_entity_contexts_from_file(self, file) -> Iterator[Tuple[str, str]]:
         """Generator that yields (entity, context) tuples from a Turtle file"""
@@ -184,6 +183,12 @@ def main():
         help=f'Collection ID (default: {default_collection})'
     )
 
+    parser.add_argument(
+        '-g', '--graph',
+        default="",
+        help='Named graph URI (e.g. urn:graph:policy). Default: default graph'
+    )
+
     parser.add_argument(
         'files', nargs='+',
         help=f'Turtle files to load'
@@ -202,6 +207,7 @@ def main():
                 files=args.files,
                 collection=args.collection,
                 workspace=args.workspace,
+                graph=args.graph,
             )
 
             loader.run()

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/load_turtle.py

@@ -9,6 +9,7 @@ import rdflib
 from typing import Iterator
 
 from trustgraph.api import Api, Triple
+from trustgraph.api.types import Uri, Literal
 from trustgraph.log_level import LogLevel
 
 default_url = os.getenv("TRUSTGRAPH_URL", 'http://localhost:8888/')
@@ -42,23 +43,19 @@ class Loader:
         g.parse(file, format="turtle")
 
         for e in g:
-            s_value = str(e[0])
-            p_value = str(e[1])
-            o_value = str(e[2])
-
-            o_datatype = ""
-            o_language = ""
+            s_value = Uri(str(e[0]))
+            p_value = Uri(str(e[1]))
 
             if isinstance(e[2], rdflib.term.Literal):
-                if e[2].language:
-                    o_language = str(e[2].language)
-                if e[2].datatype:
-                    o_datatype = str(e[2].datatype)
-
-            yield Triple(
-                s=s_value, p=p_value, o=o_value,
-                o_datatype=o_datatype, o_language=o_language,
-            )
+                o_value = Literal(
+                    str(e[2]),
+                    datatype=str(e[2].datatype) if e[2].datatype else None,
+                    language=str(e[2].language) if e[2].language else None,
+                )
+            else:
+                o_value = Uri(str(e[2]))
+
+            yield Triple(s=s_value, p=p_value, o=o_value)
 
     def run(self):
         """Load triples using Python API"""

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph/cli/nquads.py

@@ -13,7 +13,7 @@ import re
 import rdflib
 
 from trustgraph.schema import IRI, LITERAL
-from trustgraph.api.types import Triple
+from trustgraph.api.types import Triple, Uri, Literal
 
 # RDF-star quoted triples have no standard N-Quads encoding; they are
 # skipped with a count so callers can surface the omission.
@@ -39,6 +39,21 @@ _BAD_IRI = re.compile(r'[\x00-\x20<>"{}|^\x60\\]')
 _LANGTAG = re.compile(r"[a-zA-Z]+(?:-[a-zA-Z0-9]+)*")
 
 
+def valid_language_tag(language):
+    """True when language can be emitted after the '@' of a LANGTAG.
+
+    The wire schema types the language field as a free-form str, so an
+    unusable tag arrives here as data, not as a bug.
+    """
+    return isinstance(language, str) and _LANGTAG.fullmatch(language) is not None
+
+# The body of the LANGTAG production, [a-zA-Z]+ ('-' [a-zA-Z0-9]+)*, which
+# Turtle shares and which is the shape rdflib's Literal(lang=...) accepts.
+# fullmatch, not a trailing $: $ also matches before a trailing newline, so
+# "en\n" would pass and split the quad across two lines.
+_LANGTAG = re.compile(r"[a-zA-Z]+(?:-[a-zA-Z0-9]+)*")
+
+
 def valid_language_tag(language):
     """True when language can be emitted after the '@' of a LANGTAG.
 
@@ -135,22 +150,24 @@ def serialize_nquads(batches, graph_iri, out):
     return written, skipped
 
 
-def _term_to_triple_field(term):
-    """Convert an rdflib term to (str_value, datatype, language)."""
+def _rdflib_to_object(term):
+    """Convert an rdflib term to a Uri or Literal."""
     if isinstance(term, rdflib.Literal):
-        dt = str(term.datatype) if term.datatype else ""
-        lang = str(term.language) if term.language else ""
-        return str(term), dt, lang
-    return str(term), "", ""
+        return Literal(
+            str(term),
+            datatype=str(term.datatype) if term.datatype else None,
+            language=str(term.language) if term.language else None,
+        )
+    return Uri(str(term))
 
 
 def parse_nquads(data):
     """Parse N-Quads bytes back into api Triple values.
 
     Preserves datatype, language tag and IRI-vs-literal distinctions
-    via the Triple.o_datatype and Triple.o_language fields. The whole
-    member is materialized in memory (bundles are bounded by
-    --triples-limit at export); line-streaming is a possible follow-up.
+    via Literal's datatype and language attributes. The whole member is
+    materialized in memory (bundles are bounded by --triples-limit at
+    export); line-streaming is a possible follow-up.
 
     :param data: N-Quads bytes (one bundle member)
     :returns: list of Triple
@@ -159,9 +176,9 @@ def parse_nquads(data):
     ds.parse(data=data.decode("utf-8"), format="nquads")
     result = []
     for s, p, o, _g in ds.quads((None, None, None, None)):
-        o_val, o_dt, o_lang = _term_to_triple_field(o)
         result.append(Triple(
-            s=str(s), p=str(p), o=o_val,
-            o_datatype=o_dt, o_language=o_lang,
+            s=Uri(str(s)),
+            p=Uri(str(p)),
+            o=_rdflib_to_object(o),
         ))
     return result

trustgraph_cli-2.10.1/trustgraph/cli_version.py

@@ -0,0 +1 @@
+__version__ = "2.10.1"

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: trustgraph-cli
-Version: 2.9.11
+Version: 2.10.1
 Summary: TrustGraph provides a means to run a pipeline of flexible AI processing components in a flexible means to achieve a processing pipeline.
 Author-email: "trustgraph.ai" <security@trustgraph.ai>
 Project-URL: Homepage, https://github.com/trustgraph-ai/trustgraph
@@ -8,7 +8,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Operating System :: OS Independent
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
-Requires-Dist: trustgraph-base<2.10,>=2.9
+Requires-Dist: trustgraph-base<2.11,>=2.10
 Requires-Dist: requests
 Requires-Dist: pulsar-client
 Requires-Dist: aiohttp

{trustgraph_cli-2.9.11 → trustgraph_cli-2.10.1}/trustgraph_cli.egg-info/requires.txt

@@ -1,4 +1,4 @@
-trustgraph-base<2.10,>=2.9
+trustgraph-base<2.11,>=2.10
 requests
 pulsar-client
 aiohttp

trustgraph_cli-2.9.11/trustgraph/cli_version.py

@@ -1 +0,0 @@
-__version__ = "2.9.11"