Merge branch 'dpr-add-llms' of github.com:dillonroach/ragna into dpr-…

…add-llms

.github/actions/setup-env/action.yml

@@ -3,7 +3,7 @@ description: "Setup development environment"
 
 inputs:
   python-version:
-    default: "3.9"
+    default: "3.10"
     description: "Python version to install."
   optional-dependencies:
     default: "true"

.github/workflows/test.yml

@@ -40,12 +40,12 @@ jobs:
           - ubuntu-latest
           - windows-latest
           - macos-latest
-        python-version: ["3.9"]
+        python-version: ["3.10"]
         include:
-          - os: ubuntu-latest
-            python-version: "3.10"
           - os: ubuntu-latest
             python-version: "3.11"
+          - os: ubuntu-latest
+            python-version: "3.12"
 
       fail-fast: false
 
@@ -91,22 +91,22 @@ jobs:
           - chromium
           - firefox
         python-version:
-          - "3.9"
           - "3.10"
-          - "3.11"
+          - "3.10"
+          - "3.12"
         exclude:
-          - python-version: "3.10"
-            os: windows-latest
           - python-version: "3.11"
             os: windows-latest
-          - python-version: "3.10"
-            os: macos-latest
+          - python-version: "3.12"
+            os: windows-latest
           - python-version: "3.11"
             os: macos-latest
+          - python-version: "3.12"
+            os: macos-latest
         include:
           - browser: webkit
             os: macos-latest
-            python-version: "3.9"
+            python-version: "3.10"
 
       fail-fast: false
 

.github/workflows/update-docker-requirements.yml

@@ -30,7 +30,7 @@ jobs:
       - name: Setup environment
         uses: ./.github/actions/setup-env
         with:
-          python-version: "3.11"
+          python-version: "3.12"
           optional-dependencies: "false"
 
       - name: Update docker requirements

Dockerfile

@@ -1,4 +1,4 @@
-FROM python:3.11
+FROM python:3.12
 
 WORKDIR /opt/ragna
 

environment-dev.yml

@@ -2,7 +2,7 @@ name: ragna-dev
 channels:
   - conda-forge
 dependencies:
-  - python =3.9
+  - python =3.10
   - pip
   - git-lfs
   - pip:

pyproject.toml

@@ -19,13 +19,12 @@ classifiers = [
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
 ]
-requires-python = ">=3.9"
+requires-python = ">=3.10"
 dependencies = [
     "aiofiles",
     "emoji",
     "fastapi",
     "httpx",
-    "importlib_metadata>=4.6; python_version<'3.10'",
     "packaging",
     "panel==1.4.4",
     "pydantic>=2",
@@ -96,14 +95,19 @@ ignore = ["E501"]
 [tool.pytest.ini_options]
 minversion = "6.0"
 addopts = "-ra --tb=short --asyncio-mode=auto"
+asyncio_default_fixture_loop_scope = "function"
 testpaths = [
     "tests",
 ]
 filterwarnings = [
     "error",
     "ignore::ResourceWarning",
     # https://github.com/lancedb/lancedb/issues/1296
-    "ignore:Function 'semver.parse_version_info' is deprecated:DeprecationWarning"
+    "ignore:Function 'semver.parse_version_info' is deprecated:DeprecationWarning",
+    # chromadb -> opentelemetry -> google
+    "ignore:Type google._upb._message:DeprecationWarning",
+    # chromadb
+    "ignore:Python 3.14 will:DeprecationWarning"
 ]
 xfail_strict = true
 

ragna/assistants/_http_api.py

@@ -7,7 +7,6 @@
 import httpx
 
 import ragna
-from ragna._compat import anext
 from ragna.core import (
     Assistant,
     EnvVarRequirement,
@@ -132,7 +131,7 @@ async def read(self, n: int) -> bytes:
             # and set up decoding.
             if n == 0:
                 return b""
-            return await anext(self._ait, b"")  # type: ignore[call-arg]
+            return await anext(self._ait, b"")
 
     @contextlib.asynccontextmanager
     async def _stream_json(

ragna/core/_utils.py

@@ -15,11 +15,7 @@
 import pydantic
 import pydantic_core
 
-from ragna._compat import importlib_metadata_package_distributions
-
-importlib_metadata_package_distributions = functools.cache(
-    importlib_metadata_package_distributions
-)
+packages_distributions = functools.cache(importlib.metadata.packages_distributions)
 
 
 class RagnaExceptionHttpDetail(enum.Enum):
@@ -98,8 +94,8 @@ def is_available(self) -> bool:
 
         for module_name in {
             module_name
-            for module_name, distribution_names in importlib_metadata_package_distributions().items()
-            if distribution.name in distribution_names  # type: ignore[attr-defined]
+            for module_name, distribution_names in packages_distributions().items()
+            if distribution.name in distribution_names
             and module_name not in self._exclude_modules
         }:
             try:

ragna/deploy/_api/core.py

@@ -19,7 +19,6 @@
 
 import ragna
 import ragna.core
-from ragna._compat import aiter, anext
 from ragna._utils import handle_localhost_origins
 from ragna.core import Assistant, Component, Rag, RagnaException, SourceStorage
 from ragna.core._rag import SpecialChatParams

ragna/deploy/_api/orm.py

@@ -1,5 +1,6 @@
 import json
-from typing import Any
+from datetime import datetime, timezone
+from typing import Any, Optional
 
 from sqlalchemy import Column, ForeignKey, Table, types
 from sqlalchemy.engine import Dialect
@@ -30,6 +31,34 @@ def process_result_value(
         return json.loads(value)
 
 
+class UtcDateTime(types.TypeDecorator):
+    """UTC timezone aware datetime type.
+
+    This is needed because sqlalchemy.types.DateTime(timezone=True) does not
+    consistently store the timezone.
+    """
+
+    impl = types.DateTime
+
+    cache_ok = True
+
+    def process_bind_param(  # type: ignore[override]
+        self, value: Optional[datetime], dialect: Dialect
+    ) -> Optional[datetime]:
+        if value is not None:
+            assert value.tzinfo == timezone.utc
+
+        return value
+
+    def process_result_value(
+        self, value: Optional[datetime], dialect: Dialect
+    ) -> Optional[datetime]:
+        if value is None:
+            return None
+
+        return value.replace(tzinfo=timezone.utc)
+
+
 class Base(DeclarativeBase):
     pass
 
@@ -134,4 +163,4 @@ class Message(Base):
         secondary=source_message_association_table,
         back_populates="messages",
     )
-    timestamp = Column(types.DateTime, nullable=False)
+    timestamp = Column(UtcDateTime, nullable=False)

ragna/deploy/_api/schemas.py

@@ -1,14 +1,24 @@
 from __future__ import annotations
 
-import datetime
 import uuid
-from typing import Any
+from datetime import datetime, timezone
+from typing import Annotated, Any
 
-from pydantic import BaseModel, Field
+from pydantic import AfterValidator, BaseModel, Field
 
 import ragna.core
 
 
+def _set_utc_timezone(v: datetime) -> datetime:
+    if v.tzinfo is None:
+        return v.replace(tzinfo=timezone.utc)
+    else:
+        return v.astimezone(timezone.utc)
+
+
+UtcDateTime = Annotated[datetime, AfterValidator(_set_utc_timezone)]
+
+
 class Components(BaseModel):
     documents: list[str]
     source_storages: list[dict[str, Any]]
@@ -75,7 +85,7 @@ class Message(BaseModel):
     content: str
     role: ragna.core.MessageRole
     sources: list[Source] = Field(default_factory=list)
-    timestamp: datetime.datetime = Field(default_factory=datetime.datetime.utcnow)
+    timestamp: UtcDateTime = Field(default_factory=lambda: datetime.now(timezone.utc))
 
     @classmethod
     def from_core(cls, message: ragna.core.Message) -> Message:

ragna/deploy/_ui/api_wrapper.py

@@ -101,6 +101,6 @@ async def start_and_prepare(
         return chat["id"]
 
     def improve_message(self, msg):
-        msg["timestamp"] = datetime.strptime(msg["timestamp"], "%Y-%m-%dT%H:%M:%S.%f")
+        msg["timestamp"] = datetime.strptime(msg["timestamp"], "%Y-%m-%dT%H:%M:%S.%fZ")
         msg["content"] = emoji.emojize(msg["content"], language="alias")
         return msg

ragna/deploy/_ui/central_view.py

@@ -7,8 +7,6 @@
 import param
 from panel.reactive import ReactiveHTML
 
-from ragna._compat import anext
-
 from . import styles as ui
 
 

ragna/source_storages/_vector_database.py

@@ -10,7 +10,6 @@
     cast,
 )
 
-from ragna._compat import itertools_pairwise
 from ragna.core import (
     PackageRequirement,
     Page,
@@ -106,7 +105,7 @@ def _page_numbers_to_str(self, page_numbers: Optional[Iterable[int]]) -> str:
 
         ranges_str = []
         range_int = []
-        for current_page_number, next_page_number in itertools_pairwise(
+        for current_page_number, next_page_number in itertools.pairwise(
             itertools.chain(sorted(page_numbers), [None])
         ):
             current_page_number = cast(int, current_page_number)

requirements-docker.lock

@@ -1,5 +1,5 @@
 #
-# This file is autogenerated by pip-compile with Python 3.11
+# This file is autogenerated by pip-compile with Python 3.12
 # by the following command:
 #
 #    pip-compile --extra=all --output-file=requirements-docker.lock --strip-extras pyproject.toml
@@ -158,7 +158,7 @@ oauthlib==3.2.2
     # via
     #   kubernetes
     #   requests-oauthlib
-onnxruntime==1.16.3
+onnxruntime==1.19.0
     # via chromadb
 opentelemetry-api==1.22.0
     # via