Support Python 3.9 in knowledge_store and knowledge_graph (#462)

* Support Python 3.8.1

* Update ci-unit-tests.yml

* Test knowledge-store and knowledge-graph on py38 and py39

* Remove 3.8 from test matrix

* Use 3.9 as min python version

---------

Co-authored-by: Nicolò Boschi <[email protected]>

Author: cbornet

.github/workflows/ci-unit-tests.yml

@@ -79,26 +79,26 @@ jobs:
         run: tox -e unit-tests && rm -rf .tox
 
       - name: "Unit tests (colbert)"
-        if: ${{ needs.preconditions.outputs.libs_colbert == 'true'}}
+        if: ${{ needs.preconditions.outputs.libs_colbert == 'true' }}
         run: tox -e unit-tests -c libs/colbert && rm -rf libs/colbert/.tox
 
       - name: "Unit tests (langchain)"
-        if: ${{ needs.preconditions.outputs.libs_langchain == 'true'}}
+        if: ${{ needs.preconditions.outputs.libs_langchain == 'true' }}
         run: tox -e unit-tests -c libs/langchain && rm -rf libs/langchain/.tox
 
       - name: "Unit tests (llamaindex)"
-        if: ${{ needs.preconditions.outputs.libs_llamaindex == 'true'}}
+        if: ${{ needs.preconditions.outputs.libs_llamaindex == 'true' }}
         run: tox -e unit-tests -c libs/llamaindex && rm -rf libs/llamaindex/.tox
 
       - name: "Unit tests (knowledge-store)"
-        if: ${{ needs.preconditions.outputs.libs_knowledge_store == 'true' && matrix.python-version != '3.9' }}
+        if: ${{ needs.preconditions.outputs.libs_knowledge_store == 'true' }}
         env:
           OPENAI_API_KEY: "${{ secrets.E2E_TESTS_OPEN_AI_KEY }}"
         run: tox -e unit-tests -c libs/knowledge-store && rm -rf libs/knowledge-store/.tox
 
       - name: "Unit tests (knowledge-graph)"
         # yamllint disable-line rule:line-length
-        if: ${{ needs.preconditions.outputs.libs_knowledge_graph == 'true' && matrix.python-version != '3.9' && matrix.python-version != '3.10' }}
+        if: ${{ needs.preconditions.outputs.libs_knowledge_graph == 'true' }}
         env:
           OPENAI_API_KEY: "${{ secrets.E2E_TESTS_OPEN_AI_KEY }}"
         run: tox -e unit-tests -c libs/knowledge-graph && rm -rf libs/knowledge-graph/.tox

libs/colbert/pyproject.toml

@@ -10,7 +10,7 @@ documentation = "https://docs.datastax.com/en/ragstack"
 packages = [{ include = "ragstack_colbert" }]
 
 [tool.poetry.dependencies]
-python = ">=3.9,<3.13"
+python = ">=3.9,<4.0"
 colbert-ai = "0.2.19"
 pyarrow = "14.0.1"
 torch = "2.2.1"

libs/knowledge-graph/pyproject.toml

@@ -14,7 +14,7 @@ include = [
 
 
 [tool.poetry.dependencies]
-python = ">=3.11,<3.13"
+python = ">=3.8.1,<4.0"
 langchain = "^0.1.14"
 langchain-community = "^0.0.31"
 langchain-openai = "^0.1.1"
@@ -24,7 +24,6 @@ graphviz = "^0.20.3"
 pydantic-yaml = "^1.3.0"
 pyyaml = "^6.0.1"
 
-
 [tool.poetry.group.dev.dependencies]
 python-dotenv = "^1.0.1"
 ipykernel = "^6.29.4"
@@ -34,8 +33,8 @@ requests = "^2.32.2"
 pytest = "^8.1.1"
 pytest-asyncio = "^0.23.6"
 pytest-dotenv = "^0.5.2"
+pytest-rerunfailures = "^14.0"
 setuptools = "^70.0.0"
-pytest-retry = "^1.6.3"
 
 [build-system]
 requires = ["poetry-core"]

libs/knowledge-graph/ragstack_knowledge_graph/knowledge_graph.py

@@ -221,7 +221,7 @@ def insert(
 
     def subgraph(
         self,
-        start: Node | Sequence[Node],
+        start: Union[Node, Sequence[Node]],
         edge_filters: Sequence[str] = (),
         steps: int = 3,
     ) -> Tuple[Iterable[Node], Iterable[Relation]]:
@@ -249,7 +249,7 @@ def subgraph(
 
     def traverse(
         self,
-        start: Node | Sequence[Node],
+        start: Union[Node, Sequence[Node]],
         edge_filters: Sequence[str] = (),
         steps: int = 3,
     ) -> Iterable[Relation]:
@@ -280,7 +280,7 @@ def traverse(
 
     async def atraverse(
         self,
-        start: Node | Sequence[Node],
+        start: Union[Node, Sequence[Node]],
         edge_filters: Sequence[str] = (),
         steps: int = 3,
     ) -> Iterable[Relation]:

libs/knowledge-graph/ragstack_knowledge_graph/knowledge_schema.py

@@ -1,5 +1,5 @@
 from pathlib import Path
-from typing import Dict, List, Self, Sequence, Union
+from typing import Dict, List, Sequence, Union
 
 from langchain_community.graphs.graph_document import GraphDocument
 from langchain_core.pydantic_v1 import BaseModel
@@ -56,7 +56,7 @@ class KnowledgeSchema(BaseModel):
     """Allowed relationships for the knowledge schema."""
 
     @classmethod
-    def from_file(cls, path: Union[str, Path]) -> Self:
+    def from_file(cls, path: Union[str, Path]) -> "KnowledgeSchema":
         """Load a KnowledgeSchema from a JSON or YAML file.
 
         Parameters:

libs/knowledge-graph/ragstack_knowledge_graph/traverse.py

@@ -1,6 +1,6 @@
 import asyncio
 import threading
-from typing import Any, Dict, Iterable, NamedTuple, Optional, Sequence
+from typing import Any, Dict, Iterable, NamedTuple, Optional, Sequence, Union
 
 from cassandra.cluster import PreparedStatement, ResponseFuture, Session
 from cassio.config import check_resolve_keyspace, check_resolve_session
@@ -69,7 +69,7 @@ def _prepare_edge_query(
 
 
 def traverse(
-    start: Node | Sequence[Node],
+    start: Union[Node, Sequence[Node]],
     edge_table: str,
     edge_source_name: str = "source_name",
     edge_source_type: str = "source_type",
@@ -214,7 +214,7 @@ async def next(self):
 
 
 async def atraverse(
-    start: Node | Sequence[Node],
+    start: Union[Node, Sequence[Node]],
     edge_table: str,
     edge_source_name: str = "source_name",
     edge_source_type: str = "source_type",

libs/knowledge-graph/tests/test_extraction.py

@@ -34,7 +34,7 @@ def extractor(llm: BaseChatModel) -> KnowledgeSchemaExtractor:
 """
 
 
-@pytest.mark.flaky(retries=10, delay=0)
+@pytest.mark.flaky(reruns=10, reruns_delay=0)
 def test_extraction(extractor: KnowledgeSchemaExtractor):
     results = extractor.extract([Document(page_content=MARIE_CURIE_SOURCE)])
 

libs/knowledge-graph/tests/test_schema_inference.py

@@ -18,7 +18,7 @@
 """
 
 
-@pytest.mark.flaky(retries=5, delay=0)
+@pytest.mark.flaky(reruns=5, reruns_delay=0)
 def test_schema_inference(llm: BaseChatModel):
     schema_inferer = KnowledgeSchemaInferer(llm)
 

libs/knowledge-store/pyproject.toml

@@ -10,7 +10,7 @@ documentation = "https://docs.datastax.com/en/ragstack"
 packages = [{ include = "ragstack_knowledge_store" }]
 
 [tool.poetry.dependencies]
-python = ">=3.10,<3.13"
+python = ">=3.8.1,<4.0"
 langchain-core = "^0.2"
 cassio = "^0.1.7"
 

libs/knowledge-store/ragstack_knowledge_store/cassandra.py

@@ -22,7 +22,6 @@
 
 from ragstack_knowledge_store.edge_extractor import get_link_tags
 
-from ._utils import strict_zip
 from .base import KnowledgeStore, Node, TextNode
 from .concurrency import ConcurrentQueries
 from .content import Kind
@@ -331,7 +330,7 @@ def add_nodes(
 
         # Step 1: Add the nodes, collecting the tags and new sources / targets.
         with self._concurrent_queries() as cq:
-            tuples = strict_zip(texts, text_embeddings, metadatas)
+            tuples = zip(texts, text_embeddings, metadatas)
             for text, text_embedding, metadata in tuples:
                 if CONTENT_ID not in metadata:
                     metadata[CONTENT_ID] = secrets.token_hex(8)