Merge pull request #377 from ggozad/chore/deps

Update dependencies

CHANGELOG.md

@@ -1,6 +1,14 @@
 # Changelog
 ## [Unreleased]
 
+### Changed
+
+- Bump `docling>=2.93.0` and `docling-core>=2.75.0`.
+- Bump `pydantic-ai-slim>=1.96.0`. Migrate off deprecated APIs: AG-UI imports use `pydantic_ai.ui.ag_ui`, docs/CLI examples use the explicit `openai-chat:` model prefix, and `Agent(retries=)` is split into `tool_retries=` + `output_retries=`.
+- Bump `pydantic-monty>=0.0.17`. Migrate off deprecated `pydantic_monty.run_repl_async(repl, ...)` to `repl.feed_run_async(...)`.
+- Cap `transformers<5.0.0` in the `mxbai` extra: `mxbai-rerank>=0.1.6` calls `tokenizer.prepare_for_model` which transformers 5 removed.
+- Refresh the rest of the lockfile to latest within current constraints (pydantic, pydantic-ai, rich, ruff, ty, pytest, torch, textual, textual-image, watchfiles, pre-commit, datasets, and transitives).
+
 ## [0.47.0] - 2026-05-14
 
 ### Added

app/backend/main.py

@@ -6,8 +6,8 @@ from pathlib import Path
 
 from dotenv import find_dotenv, load_dotenv
 from pydantic_ai import Agent
-from pydantic_ai.ag_ui import AGUIAdapter
 from pydantic_ai.ui import SSE_CONTENT_TYPE
+from pydantic_ai.ui.ag_ui import AGUIAdapter
 from starlette.applications import Starlette
 from starlette.middleware import Middleware
 from starlette.middleware.cors import CORSMiddleware

docs/python.md

@@ -490,7 +490,7 @@ skill = create_skill(db_path=db_path, config=config)
 toolset = SkillToolset(skills=[skill])
 
 agent = Agent(
-    "openai:gpt-4o",
+    "openai-chat:gpt-4o",
     instructions=build_system_prompt(toolset.skill_catalog),
     toolsets=[toolset],
 )

docs/skills/analysis.md

@@ -59,7 +59,7 @@ analysis = create_analysis_skill(db_path=db_path)
 toolset = SkillToolset(skills=[rag, analysis])
 
 agent = Agent(
-    "openai:gpt-4o",
+    "openai-chat:gpt-4o",
     instructions=build_system_prompt(toolset.skill_catalog),
     toolsets=[toolset],
 )

docs/skills/index.md

@@ -31,7 +31,7 @@ skill = create_skill(db_path=db_path, config=config)
 toolset = SkillToolset(skills=[skill])
 
 agent = Agent(
-    "openai:gpt-4o",
+    "openai-chat:gpt-4o",
     instructions=build_system_prompt(toolset.skill_catalog),
     toolsets=[toolset],
 )
@@ -98,7 +98,7 @@ See the individual skill pages for state model details.
 For web applications, use pydantic-ai's `AGUIAdapter` to stream tool calls, text, and state deltas:
 
 ```python
-from pydantic_ai.ag_ui import AGUIAdapter
+from pydantic_ai.ui.ag_ui import AGUIAdapter
 
 adapter = AGUIAdapter(agent=agent, run_input=run_input)
 event_stream = adapter.run_stream()

evaluations/evaluations/evaluators/judge.py

@@ -51,7 +51,8 @@ class LLMJudge:
             model=model_obj,
             output_type=LLMJudgeResponseSchema,
             system_prompt=ANSWER_EQUIVALENCE_RUBRIC,
-            retries=3,
+            tool_retries=3,
+            output_retries=3,
         )
 
     async def judge_answers(

examples/custom_agent_agui.py

@@ -16,8 +16,8 @@ import sys
 from pathlib import Path
 
 from pydantic_ai import Agent
-from pydantic_ai.ag_ui import AGUIAdapter
 from pydantic_ai.ui import SSE_CONTENT_TYPE
+from pydantic_ai.ui.ag_ui import AGUIAdapter
 from starlette.applications import Starlette
 from starlette.requests import Request
 from starlette.responses import JSONResponse, Response, StreamingResponse

haiku_rag_slim/haiku/rag/agents/analysis/agent.py

@@ -27,7 +27,8 @@ def create_analysis_agent(config: AppConfig) -> Agent[AnalysisDeps, RawAnalysisR
         deps_type=AnalysisDeps,
         output_type=RawAnalysisResult,
         instructions=ANALYSIS_SYSTEM_PROMPT,
-        retries=3,
+        tool_retries=3,
+        output_retries=3,
     )
 
     @agent.tool

haiku_rag_slim/haiku/rag/agents/analysis/sandbox.py

@@ -260,8 +260,7 @@ class Sandbox:
                 },
             )
             if self._context.documents:
-                await pydantic_monty.run_repl_async(
+                await self._repl.feed_run_async(
-                    self._repl,
                     "pass",
                     inputs={
                         "documents": [
@@ -296,8 +295,7 @@ class Sandbox:
         max_chars = self._config.analysis.max_output_chars
 
         try:
-            output = await pydantic_monty.run_repl_async(
+            output = await repl.feed_run_async(
-                repl,
                 code,
                 external_functions=external_fns,
                 print_callback=print_callback,

haiku_rag_slim/haiku/rag/agents/qa/agent.py

@@ -69,7 +69,8 @@ class QuestionAnswerAgent:
             output_type=RawSearchAnswer,
             instructions=system_prompt,
             toolsets=[search_toolset],
-            retries=3,
+            tool_retries=3,
+            output_retries=3,
         )
 
         deps = _QARunDeps(client=self._client)

haiku_rag_slim/haiku/rag/agents/research/graph.py

@@ -70,7 +70,8 @@ async def _iterative_plan_logic(
         model=model,
         output_type=IterativePlanResult,
         instructions=effective_prompt,
-        retries=3,
+        tool_retries=3,
+        output_retries=3,
         deps_type=ResearchDependencies,
     )
 
@@ -120,7 +121,8 @@ async def _search_one_step_logic(
             model=model,
             output_type=RawSearchAnswer,
             instructions=search_prompt,
-            retries=3,
+            tool_retries=3,
+            output_retries=3,
             deps_type=ResearchDependencies,
         )
 
@@ -222,7 +224,8 @@ def build_research_graph(
             model=model,
             output_type=ResearchReport,
             instructions=synthesis_prompt,
-            retries=3,
+            tool_retries=3,
+            output_retries=3,
             deps_type=ResearchDependencies,
         )
 

haiku_rag_slim/haiku/rag/chat/app.py

@@ -21,7 +21,7 @@ from ag_ui.core import (
 )
 from jsonpatch import JsonPatch
 from pydantic_ai import Agent
-from pydantic_ai.ag_ui import AGUIAdapter
+from pydantic_ai.ui.ag_ui import AGUIAdapter
 from textual.app import App, SystemCommand
 from textual.binding import Binding
 from textual.widgets import Footer, Header, Input

haiku_rag_slim/haiku/rag/cli.py

@@ -663,7 +663,7 @@ def chat(  # pragma: no cover
     model: str | None = typer.Option(
         None,
         "--model",
-        help="Model to use for the chat (e.g. openai:gpt-4o)",
+        help="Model to use for the chat (e.g. openai-chat:gpt-4o)",
     ),
     skill: list[str] | None = typer.Option(
         None,

haiku_rag_slim/haiku/rag/client/rebuild.py

@@ -393,6 +393,8 @@ async def _rebuild_embed_only(
             embeddings.extend(batch_embeddings)
 
         for chunk, content_fts, embedding in zip(chunks, texts, embeddings):
+            assert chunk.id is not None
+            assert chunk.document_id is not None
             pending_records.append(
                 client.store.ChunkRecord(
                     id=chunk.id,

haiku_rag_slim/haiku/rag/store/upgrades/v0_25_0.py

@@ -96,7 +96,7 @@ async def _apply_compress_docling_document(store: Store) -> None:  # pragma: no
                     "Recovering %d documents from failed migration", len(staging_ids)
                 )
                 # Create new documents table and copy from staging
-                store.documents_table = None
+                del store.documents_table
                 if "documents" in (await store.db.list_tables()).tables:
                     await store.db.drop_table("documents")
                 store.documents_table = await store.db.create_table(
@@ -137,7 +137,7 @@ async def _apply_compress_docling_document(store: Store) -> None:  # pragma: no
                 return
 
         # No documents and no staging to recover, just recreate table with new schema
-        store.documents_table = None
+        del store.documents_table
         if "documents" in (await store.db.list_tables()).tables:
             await store.db.drop_table("documents")
         store.documents_table = await store.db.create_table(
@@ -177,7 +177,7 @@ async def _apply_compress_docling_document(store: Store) -> None:  # pragma: no
         )
 
     # Replace old table with staging table
-    store.documents_table = None
+    del store.documents_table
     if "documents" in (await store.db.list_tables()).tables:
         await store.db.drop_table("documents")
     store.documents_table = await store.db.create_table(

haiku_rag_slim/haiku/rag/store/upgrades/v0_38_0.py

@@ -118,7 +118,7 @@ async def _apply_split_pages_zstd(store: Store) -> None:  # pragma: no cover
                 logger.info(
                     "Recovering %d documents from failed migration", len(staging_ids)
                 )
-                store.documents_table = None
+                del store.documents_table
                 if "documents" in (await store.db.list_tables()).tables:
                     await store.db.drop_table("documents")
                 store.documents_table = await store.db.create_table(
@@ -144,7 +144,7 @@ async def _apply_split_pages_zstd(store: Store) -> None:  # pragma: no cover
                 return
 
         # No documents and no staging — recreate table with new schema
-        store.documents_table = None
+        del store.documents_table
         if "documents" in (await store.db.list_tables()).tables:
             await store.db.drop_table("documents")
         store.documents_table = await store.db.create_table(
@@ -188,7 +188,7 @@ async def _apply_split_pages_zstd(store: Store) -> None:  # pragma: no cover
         )
 
     # Replace old table with staging table
-    store.documents_table = None
+    del store.documents_table
     if "documents" in (await store.db.list_tables()).tables:
         await store.db.drop_table("documents")
     store.documents_table = await store.db.create_table(

haiku_rag_slim/pyproject.toml

@@ -22,7 +22,7 @@ classifiers = [
 ]
 
 dependencies = [
-    "docling-core>=2.74.1",
+    "docling-core>=2.75.0",
     "haiku.skills>=0.16.0",
     "httpx>=0.28.1",
     "jinja2>=3.1.0",
@@ -30,8 +30,8 @@ dependencies = [
     "lancedb==0.30.2",
     "pathspec>=1.0.4",
     "pydantic>=2.12.5",
-    "pydantic-ai-slim[openai,fastmcp,logfire,ag-ui]>=1.81.0",
+    "pydantic-ai-slim[openai,fastmcp,logfire,ag-ui]>=1.96.0",
-    "pydantic-monty>=0.0.9",
+    "pydantic-monty>=0.0.17",
     "python-dotenv>=1.2.2",
     "pyyaml>=6.0.3",
     "rich>=14.3.3",
@@ -42,13 +42,13 @@ dependencies = [
 
 [project.optional-dependencies]
 # Document processing
-docling = ["docling>=2.91.0", "opencv-python-headless>=4.13.0.92"]
+docling = ["docling>=2.93.0", "opencv-python-headless>=4.13.0.92"]
 # S3 / object-storage monitoring
 s3 = ["obstore>=0.9,<0.10"]
 # Embedding providers
 voyageai = ["pydantic-ai-slim[voyageai]"]
 # Rerankers
-mxbai = ["mxbai-rerank>=0.1.6"]
+mxbai = ["mxbai-rerank>=0.1.6", "transformers>=4.49.0,<5.0.0"]
 cohere = ["cohere>=5.21.1"]
 zeroentropy = ["zeroentropy>=0.1.0a11"]
 jina = ["transformers>=4.40.0", "torch>=2.0.0"]