veris-ai 1.4.0__tar.gz → 1.6.0__tar.gz

{veris_ai-1.4.0 → veris_ai-1.6.0}/.cursor/rules/documentation-management.mdc

@@ -4,11 +4,12 @@ alwaysApply: false
 ---
 
 # Documentation Handling Guidelines
+Ensure READMEs function as both a human-friendly guide and an LLM semantic router, while designating source code as the sole source of truth. This document also defines LLM workflows that continuously evolve documentation in response to actual usage patterns.
 
 ### Dual‑Use Overview
 
 * **Humans**: READMEs must be narrative‑driven, intuitive, and structured for readability, with inline cross‑links to related READMEs or code modules for deeper context.
-* **LLM Agents**: Treat these same READMEs as semantic routers. Use linked references and tags within the document to locate the most relevant code, tests, and workflows — never treat the README as the ultimate logic source.
+* **LLM Agents**: Treat these same READMEs as semantic routers and exploration waypoints. Use linked references and tags to dynamically navigate through the documentation web, following trails of dependency and discovering emergent patterns — never treat the README as the ultimate logic source.
 * **Shared Goal**: Documentation must actively fight codebase complexity. Instead of growing endlessly, it should simplify, subtract redundancy, and delegate details to lower‑level READMEs in the hierarchy.
 
 ---
@@ -33,6 +34,7 @@ Updates should simplify where possible: remove outdated or redundant content, an
 2. **Parallel instance memory**: Maintain awareness of session context across LLM instances to keep documentation aligned with ongoing workflows.
 3. **LLM as thought partner**: Propose not only wording edits but also simplifications and delegation opportunities — e.g., linking to an existing module README rather than duplicating logic.
 4. **Complexity management**: Treat every update as a chance to prune. The README should remain a high‑level, navigable entry point, not a catch‑all.
+5. **Recursive documentation weaving**: When creating or updating a README, treat it as the entry point to a documentation web. Draft the initial README, then follow every dependency thread—navigating to parent and child READMEs, creating missing pieces, and refining connections. Continue this cyclical exploration until the entire documentation network achieves seamless consistency, with each iteration strengthening the coherence of the whole system.
 
 ---
 
@@ -41,13 +43,14 @@ Updates should simplify where possible: remove outdated or redundant content, an
 Each meaningful workspace or module must include a `README.md` designed to operate like a **hub page** in an IDE‑backed website:
 
 * **Quick Reference**: Purpose, setup, usage, and high‑level architecture.
-* **Linked Context**: Cross‑links to deeper READMEs, design docs, or code directories.
-* **Semantic Anchors**: Inline cues (e.g., tags, headings, links) that help the LLM map concepts to code without requiring redundant prose.
+* **Linked Context**: Cross‑links to deeper child READMEs and parent READMEs, design docs, or code directories.
+* **Visual Aids**: Use visual aids (e.g. mermaid diagrams) to as needed to help explain the codebase to humans.
+
 
 > Example:
 > *“For transaction processing, see [Transactions README](./transactions/README.md). For error handling logic, see [Error Handling README](./errors/README.md).”*
 
-The human and LLM share the same document: humans follow the narrative, while the LLM uses references and anchors to navigate the codebase semantically.
+The human and LLM share the same document: humans follow the narrative, while the LLM uses references in the LLM to navigate the codebase semantically.
 
 ---
 
@@ -57,12 +60,11 @@ The human and LLM share the same document: humans follow the narrative, while th
 
 * Detect README drift by comparing live code to described behavior.
 * Perform updates with an emphasis on pruning duplication and linking to existing READMEs.
-* Use end‑of‑session summaries to suggest or implement simplifications.
 * Ensure docs remain aligned with code without ballooning in size.
 
 #### Human Responsibilities
 
-* Review LLM‑driven updates for clarity, accuracy, and usability.
+* Diligently read through *ALL* LLM‑driven documentation updates for clarity, accuracy, and usability.
 * Refactor prose when needed to keep explanations intuitive.
 * Validate that cross‑links resolve correctly and are helpful for navigation.
 
@@ -77,4 +79,4 @@ The human and LLM share the same document: humans follow the narrative, while th
 | **Simplicity over sprawl** | Fight complexity by pruning, delegating, and cross‑linking.               |
 | **One README, two roles**  | The same README serves both humans and LLMs through cross‑referencing.    |
 | **Real‑world grounding**   | Updates reflect actual changes in workflows and architecture.             |
-| **Human validation**       | LLM edits require human review to ensure usability and accuracy.          |
+| **Human validation**       | LLM documentation edits require thorough human driver review to ensure usability and accuracy.          |

veris_ai-1.6.0/.pre-commit-config.yaml

@@ -0,0 +1,24 @@
+repos:
+  - repo: local
+    hooks:
+      - id: ruff-check
+        name: Ruff lint
+        entry: uv run ruff check .
+        language: system
+        pass_filenames: false
+        types: [python]
+
+      - id: ruff-format-check
+        name: Ruff format (check)
+        entry: uv run ruff format --check .
+        language: system
+        pass_filenames: false
+        types: [python]
+
+      - id: mypy
+        name: Mypy type check
+        entry: uv run mypy src/veris_ai tests
+        language: system
+        pass_filenames: false
+        types: [python]
+

{veris_ai-1.4.0 → veris_ai-1.6.0}/CHANGELOG.md

@@ -5,20 +5,6 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
-
-### Added
-- Session ID tracking in OTEL traces: The `veris.session_id` attribute is now automatically added to all spans when a session ID is present
-  - Session IDs from FastAPI MCP bearer tokens are automatically embedded in traces
-  - Manual session management via `veris.set_session_id()` and `veris.clear_session_id()` 
-  - Enables filtering and correlation of traces by user session in Jaeger
-- New test suite for session ID tracing functionality
-
-### Changed
-- `AgentsOTELBridgeProcessor` now imports and uses `_session_id_context` from `tool_mock`
-- Trace and span creation now includes session ID attributes when available
-
-
 ## v0.2.1 (2025-04-18)
 
 ### Bug Fixes

{veris_ai-1.4.0 → veris_ai-1.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: veris-ai
-Version: 1.4.0
+Version: 1.6.0
 Summary: A Python package for Veris AI tools
 Project-URL: Homepage, https://github.com/veris-ai/veris-python-sdk
 Project-URL: Bug Tracker, https://github.com/veris-ai/veris-python-sdk/issues
@@ -9,8 +9,17 @@ License-Expression: MIT
 License-File: LICENSE
 Requires-Python: >=3.11
 Requires-Dist: httpx>=0.24.0
+Requires-Dist: opentelemetry-api>=1.34.1
+Requires-Dist: opentelemetry-exporter-otlp>=1.34.1
+Requires-Dist: opentelemetry-instrumentation-fastapi>=0.55b1
+Requires-Dist: opentelemetry-instrumentation-httpx>=0.55b1
+Requires-Dist: opentelemetry-instrumentation-mcp>=0.44.1
+Requires-Dist: opentelemetry-instrumentation-requests>=0.55b1
+Requires-Dist: opentelemetry-instrumentation>=0.55b1
+Requires-Dist: opentelemetry-sdk>=1.34.1
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: requests>=2.31.0
+Requires-Dist: traceloop-sdk>=0.45.4
 Provides-Extra: dev
 Requires-Dist: black>=23.7.0; extra == 'dev'
 Requires-Dist: mypy>=1.5.1; extra == 'dev'
@@ -22,14 +31,9 @@ Requires-Dist: pytest>=7.4.0; extra == 'dev'
 Requires-Dist: ruff>=0.11.4; extra == 'dev'
 Provides-Extra: fastapi
 Requires-Dist: fastapi; extra == 'fastapi'
-Requires-Dist: fastapi-mcp; extra == 'fastapi'
+Requires-Dist: fastapi-mcp>=0.4.0; extra == 'fastapi'
 Provides-Extra: instrument
-Requires-Dist: braintrust; extra == 'instrument'
 Requires-Dist: opentelemetry-api; extra == 'instrument'
-Requires-Dist: opentelemetry-exporter-otlp; extra == 'instrument'
-Requires-Dist: opentelemetry-exporter-otlp-proto-common; extra == 'instrument'
-Requires-Dist: opentelemetry-exporter-otlp-proto-grpc; extra == 'instrument'
-Requires-Dist: opentelemetry-exporter-otlp-proto-http; extra == 'instrument'
 Requires-Dist: opentelemetry-sdk; extra == 'instrument'
 Requires-Dist: wrapt; extra == 'instrument'
 Description-Content-Type: text/markdown
@@ -41,7 +45,7 @@ A Python package for Veris AI tools with simulation capabilities and FastAPI MCP
 ## Quick Reference
 
 **Purpose**: Tool mocking, tracing, and FastAPI MCP integration for AI agent development  
-**Core Components**: [`tool_mock`](#function-mocking) • [`jaeger_interface`](#jaeger-trace-interface) • [`braintrust_tracing`](#tracing-integration) • [`fastapi_mcp`](#fastapi-mcp-integration)  
+**Core Components**: [`tool_mock`](#function-mocking) • [`observability`](#sdk-observability-helpers) • [`fastapi_mcp`](#fastapi-mcp-integration) • [`jaeger_interface`](#jaeger-trace-interface)  
 **Deep Dive**: [`Module Architecture`](src/veris_ai/README.md) • [`Testing Guide`](tests/README.md) • [`Usage Examples`](examples/README.md)  
 **Source of Truth**: Implementation details in [`src/veris_ai/`](src/veris_ai/) source code
 
@@ -58,7 +62,7 @@ uv add "veris-ai[dev,fastapi,instrument]"
 **Installation Profiles**:
 - `dev`: Development tools (ruff, pytest, mypy) 
 - `fastapi`: FastAPI MCP integration
-- `instrument`: Braintrust/OpenTelemetry tracing
+- `observability`: OpenTelemetry tracing
 
 ## Import Patterns
 
@@ -69,7 +73,7 @@ uv add "veris-ai[dev,fastapi,instrument]"
 from veris_ai import veris, JaegerClient
 
 # Optional features (require extras)
-from veris_ai import braintrust_tracing  # Requires [instrument]
+from veris_ai import init_observability, instrument_fastapi_app  # Provided by SDK observability helpers
 ```
 
 **Complete Import Strategies**: See [`examples/README.md`](examples/README.md) for five different import approaches, conditional features, and integration patterns.
@@ -88,22 +92,33 @@ from veris_ai import braintrust_tracing  # Requires [instrument]
 
 **Configuration Details**: See [`src/veris_ai/tool_mock.py`](src/veris_ai/tool_mock.py) for environment handling logic.
 
-## Tracing Integration
 
-**Semantic Tag**: `distributed-tracing`
+### SDK Observability Helpers
 
-Parallel tracing to Braintrust and Jaeger/OpenTelemetry for monitoring and evaluation.
+The SDK provides optional-safe observability helpers that standardize OpenTelemetry setup and W3C context propagation across services.
 
 ```python
-from veris_ai import braintrust_tracing
+from fastapi import FastAPI
+from veris_ai import init_observability, instrument_fastapi_app
+
+# Initialize tracing/export early (no-op if dependencies are absent)
+init_observability(service_name="my-customer-service")
+
+app = FastAPI()
 
-# Enable dual tracing
-braintrust_tracing.instrument(service_name="my-service", otlp_endpoint="http://localhost:4317")
+# Ensure inbound HTTP requests continue W3C traces
+instrument_fastapi_app(app)
 ```
 
-**Session Management**: Automatic session ID extraction from bearer tokens. Manual session control via `veris.set_session_id()` and `veris.clear_session_id()`.
+What this enables:
+- Sets global W3C propagator (TraceContext + Baggage)
+- Optionally instruments FastAPI, requests, httpx, MCP client if installed
+- Includes request hooks to attach outbound `traceparent` on HTTP calls for continuity
 
-**Implementation Details**: See [`src/veris_ai/braintrust_tracing.py`](src/veris_ai/braintrust_tracing.py) for instrumentation logic.
+End-to-end propagation with the simulator:
+- The simulator injects W3C headers when connecting to your FastAPI MCP endpoints
+- The SDK injects W3C headers on `/api/v2/tool_mock` and logging requests back to the simulator
+- Result: customer agent spans and tool mocks appear under the same distributed trace
 
 ## Function Mocking
 
@@ -139,7 +154,7 @@ async def get_data() -> dict:
 
 **Semantic Tag**: `fastapi-mcp`
 
-Expose FastAPI endpoints as MCP tools for AI agent consumption.
+Expose FastAPI endpoints as MCP tools for AI agent consumption using HTTP transport.
 
 ```python
 from fastapi import FastAPI
@@ -147,21 +162,27 @@ from veris_ai import veris
 
 app = FastAPI()
 
-# Enable MCP integration
+# Enable MCP integration with HTTP transport
 veris.set_fastapi_mcp(
     fastapi=app,
     name="My API Server",
     include_operations=["get_users", "create_user"],
     exclude_tags=["internal"]
 )
+
+# Mount the MCP server with HTTP transport (recommended)
+veris.fastapi_mcp.mount_http()
 ```
 
 **Key Features**:
+- **HTTP Transport**: Uses Streamable HTTP protocol for better session management
 - **Automatic schema conversion**: FastAPI OpenAPI → MCP tool definitions
 - **Session management**: Bearer token → session ID mapping
 - **Filtering**: Include/exclude operations and tags
 - **Authentication**: OAuth2 integration
 
+**Transport Protocol**: The SDK uses HTTP transport (via `mount_http()`) which implements the MCP Streamable HTTP specification, providing robust connection handling and fixing session routing issues with concurrent connections.
+
 **Configuration Reference**: See function signature in [`src/veris_ai/tool_mock.py`](src/veris_ai/tool_mock.py) for all `set_fastapi_mcp()` parameters.
 
 ## Utility Functions
@@ -201,7 +222,7 @@ pytest --cov=veris_ai # Test with coverage
 
 **Semantic Tag**: `module-architecture`
 
-**Core Modules**: `tool_mock` (mocking), `jaeger_interface` (trace queries), `braintrust_tracing` (dual tracing), `utils` (schema conversion)
+**Core Modules**: `tool_mock` (mocking), `jaeger_interface` (trace queries), `utils` (schema conversion)
 
 **Complete Architecture**: See [`src/veris_ai/README.md`](src/veris_ai/README.md) for module overview, implementation flows, and configuration details. 
 

{veris_ai-1.4.0 → veris_ai-1.6.0}/README.md

@@ -5,7 +5,7 @@ A Python package for Veris AI tools with simulation capabilities and FastAPI MCP
 ## Quick Reference
 
 **Purpose**: Tool mocking, tracing, and FastAPI MCP integration for AI agent development  
-**Core Components**: [`tool_mock`](#function-mocking) • [`jaeger_interface`](#jaeger-trace-interface) • [`braintrust_tracing`](#tracing-integration) • [`fastapi_mcp`](#fastapi-mcp-integration)  
+**Core Components**: [`tool_mock`](#function-mocking) • [`observability`](#sdk-observability-helpers) • [`fastapi_mcp`](#fastapi-mcp-integration) • [`jaeger_interface`](#jaeger-trace-interface)  
 **Deep Dive**: [`Module Architecture`](src/veris_ai/README.md) • [`Testing Guide`](tests/README.md) • [`Usage Examples`](examples/README.md)  
 **Source of Truth**: Implementation details in [`src/veris_ai/`](src/veris_ai/) source code
 
@@ -22,7 +22,7 @@ uv add "veris-ai[dev,fastapi,instrument]"
 **Installation Profiles**:
 - `dev`: Development tools (ruff, pytest, mypy) 
 - `fastapi`: FastAPI MCP integration
-- `instrument`: Braintrust/OpenTelemetry tracing
+- `observability`: OpenTelemetry tracing
 
 ## Import Patterns
 
@@ -33,7 +33,7 @@ uv add "veris-ai[dev,fastapi,instrument]"
 from veris_ai import veris, JaegerClient
 
 # Optional features (require extras)
-from veris_ai import braintrust_tracing  # Requires [instrument]
+from veris_ai import init_observability, instrument_fastapi_app  # Provided by SDK observability helpers
 ```
 
 **Complete Import Strategies**: See [`examples/README.md`](examples/README.md) for five different import approaches, conditional features, and integration patterns.
@@ -52,22 +52,33 @@ from veris_ai import braintrust_tracing  # Requires [instrument]
 
 **Configuration Details**: See [`src/veris_ai/tool_mock.py`](src/veris_ai/tool_mock.py) for environment handling logic.
 
-## Tracing Integration
 
-**Semantic Tag**: `distributed-tracing`
+### SDK Observability Helpers
 
-Parallel tracing to Braintrust and Jaeger/OpenTelemetry for monitoring and evaluation.
+The SDK provides optional-safe observability helpers that standardize OpenTelemetry setup and W3C context propagation across services.
 
 ```python
-from veris_ai import braintrust_tracing
+from fastapi import FastAPI
+from veris_ai import init_observability, instrument_fastapi_app
+
+# Initialize tracing/export early (no-op if dependencies are absent)
+init_observability(service_name="my-customer-service")
+
+app = FastAPI()
 
-# Enable dual tracing
-braintrust_tracing.instrument(service_name="my-service", otlp_endpoint="http://localhost:4317")
+# Ensure inbound HTTP requests continue W3C traces
+instrument_fastapi_app(app)
 ```
 
-**Session Management**: Automatic session ID extraction from bearer tokens. Manual session control via `veris.set_session_id()` and `veris.clear_session_id()`.
+What this enables:
+- Sets global W3C propagator (TraceContext + Baggage)
+- Optionally instruments FastAPI, requests, httpx, MCP client if installed
+- Includes request hooks to attach outbound `traceparent` on HTTP calls for continuity
 
-**Implementation Details**: See [`src/veris_ai/braintrust_tracing.py`](src/veris_ai/braintrust_tracing.py) for instrumentation logic.
+End-to-end propagation with the simulator:
+- The simulator injects W3C headers when connecting to your FastAPI MCP endpoints
+- The SDK injects W3C headers on `/api/v2/tool_mock` and logging requests back to the simulator
+- Result: customer agent spans and tool mocks appear under the same distributed trace
 
 ## Function Mocking
 
@@ -103,7 +114,7 @@ async def get_data() -> dict:
 
 **Semantic Tag**: `fastapi-mcp`
 
-Expose FastAPI endpoints as MCP tools for AI agent consumption.
+Expose FastAPI endpoints as MCP tools for AI agent consumption using HTTP transport.
 
 ```python
 from fastapi import FastAPI
@@ -111,21 +122,27 @@ from veris_ai import veris
 
 app = FastAPI()
 
-# Enable MCP integration
+# Enable MCP integration with HTTP transport
 veris.set_fastapi_mcp(
     fastapi=app,
     name="My API Server",
     include_operations=["get_users", "create_user"],
     exclude_tags=["internal"]
 )
+
+# Mount the MCP server with HTTP transport (recommended)
+veris.fastapi_mcp.mount_http()
 ```
 
 **Key Features**:
+- **HTTP Transport**: Uses Streamable HTTP protocol for better session management
 - **Automatic schema conversion**: FastAPI OpenAPI → MCP tool definitions
 - **Session management**: Bearer token → session ID mapping
 - **Filtering**: Include/exclude operations and tags
 - **Authentication**: OAuth2 integration
 
+**Transport Protocol**: The SDK uses HTTP transport (via `mount_http()`) which implements the MCP Streamable HTTP specification, providing robust connection handling and fixing session routing issues with concurrent connections.
+
 **Configuration Reference**: See function signature in [`src/veris_ai/tool_mock.py`](src/veris_ai/tool_mock.py) for all `set_fastapi_mcp()` parameters.
 
 ## Utility Functions
@@ -165,7 +182,7 @@ pytest --cov=veris_ai # Test with coverage
 
 **Semantic Tag**: `module-architecture`
 
-**Core Modules**: `tool_mock` (mocking), `jaeger_interface` (trace queries), `braintrust_tracing` (dual tracing), `utils` (schema conversion)
+**Core Modules**: `tool_mock` (mocking), `jaeger_interface` (trace queries), `utils` (schema conversion)
 
 **Complete Architecture**: See [`src/veris_ai/README.md`](src/veris_ai/README.md) for module overview, implementation flows, and configuration details. 
 

{veris_ai-1.4.0 → veris_ai-1.6.0}/examples/README.md

@@ -47,23 +47,16 @@ from veris_ai.tool_mock import veris as mock_tool
 **Use Case**: Maximum control over imported dependencies  
 **Benefits**: Explicit imports, reduced memory footprint
 
-### 4. Conditional Environment-Based Imports
-```python
-if os.getenv("ENABLE_TRACING") == "true":
-    from veris_ai.braintrust_tracing import instrument
-```
-**Use Case**: Feature flags, environment-specific behavior  
-**Pattern**: Runtime import decisions based on configuration
-
-### 5. FastAPI Integration
+### 4. FastAPI Integration with HTTP Transport
 ```python
 if os.getenv("USE_FASTAPI") == "true":
     from fastapi import FastAPI
     app = FastAPI()
     veris.set_fastapi_mcp(fastapi=app, name="My API Server")
+    veris.fastapi_mcp.mount_http()  # Mount with HTTP transport
 ```
-**Use Case**: MCP server setup with conditional FastAPI integration  
-**Dependencies**: Requires `[fastapi]` extra
+**Use Case**: MCP server setup with HTTP transport for robust session management  
+**Dependencies**: Requires `[fastapi]` extra (fastapi-mcp>=0.4.0)
 
 ## Integration Patterns
 

{veris_ai-1.4.0 → veris_ai-1.6.0}/examples/import_options.py

@@ -50,18 +50,7 @@ def health_check() -> dict[str, str]:
     return {"status": "healthy", "timestamp": "2024-01-01"}
 
 
-# Option 4: Conditional imports based on environment
-
-if os.getenv("ENABLE_TRACING") == "true":
-    try:
-        from veris_ai.braintrust_tracing import instrument
-
-        instrument()
-    except ImportError:
-        print("Tracing requested but dependencies not installed")
-
-
-# Option 5: FastAPI integration (requires [fastapi] extra)
+# Option 4: FastAPI integration (requires [fastapi] extra)
 if os.getenv("USE_FASTAPI") == "true":
     try:
         from fastapi import FastAPI
@@ -73,5 +62,8 @@ if os.getenv("USE_FASTAPI") == "true":
             fastapi=app,
             name="My API Server",
         )
+        # Mount the MCP server with HTTP transport (recommended)
+        # Note: User must call this separately after set_fastapi_mcp
+        veris.fastapi_mcp.mount_http()
     except ImportError:
         print("FastAPI integration requires: pip install veris-ai[fastapi]")

{veris_ai-1.4.0 → veris_ai-1.6.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "veris-ai"
-version = "1.4.0"
+version = "1.6.0"
 description = "A Python package for Veris AI tools"
 readme = "README.md"
 requires-python = ">=3.11"
@@ -16,6 +16,15 @@ dependencies = [
     "httpx>=0.24.0",
     "pydantic>=2.0.0",
     "requests>=2.31.0",
+    "opentelemetry-sdk>=1.34.1",
+    "opentelemetry-exporter-otlp>=1.34.1",
+    "opentelemetry-instrumentation>=0.55b1",
+    "opentelemetry-instrumentation-fastapi>=0.55b1",
+    "opentelemetry-instrumentation-requests>=0.55b1",
+    "opentelemetry-instrumentation-httpx>=0.55b1",
+    "opentelemetry-instrumentation-mcp>=0.44.1",
+    "traceloop-sdk>=0.45.4",
+    "opentelemetry-api>=1.34.1",
 ]
 
 [project.optional-dependencies]
@@ -31,17 +40,12 @@ dev = [
 ]
 fastapi = [
     "fastapi",
-    "fastapi_mcp"
+    "fastapi-mcp>=0.4.0",
 ]
 instrument = [
-    "braintrust",
     "wrapt",
     "opentelemetry-api",
     "opentelemetry-sdk",
-    "opentelemetry-exporter-otlp",
-    "opentelemetry-exporter-otlp-proto-grpc",
-    "opentelemetry-exporter-otlp-proto-http",
-    "opentelemetry-exporter-otlp-proto-common"
 ]
 
 [project.urls]
@@ -103,6 +107,13 @@ ignore = [
     "ANN204",  # Missing type annotation for *args
     "ANN205",  # Missing type annotation for **kwargs
     "COM812", # Trailing comma conflicts with formatter
+    "PLC0415",  # allow non-top-level imports
+    "I001",  # ignore unsorted/unformatted import block
+    "EM101", # ignore exception must not use a string literal
+    "D202", # ignore No blank lines allowed after function docstring
+    "C901", # ignore Complexity is too high
+    "TC002", # ignore move X into type checking
+    "PLR0915", # ignore Too many statements
 ]
 
 [tool.ruff.lint.isort]
@@ -153,3 +164,16 @@ tag_format = "v{version}"
 
 [tool.uv.sources]
 veris-ai = { workspace = true }
+
+[dependency-groups]
+dev = [
+    "fastapi>=0.116.1",
+    "fastapi-mcp>=0.4.0",
+    "mypy>=1.17.0",
+    "openai-agents>=0.0.1",
+    "pytest>=8.4.1",
+    "pytest-asyncio>=1.1.0",
+    "pytest-cov>=6.2.1",
+    "ruff>=0.12.4",
+    "types-requests>=2.32.4.20250611",
+]

{veris_ai-1.4.0 → veris_ai-1.6.0}/src/veris_ai/README.md

@@ -15,7 +15,6 @@ This module contains the core implementation of the Veris AI Python SDK. Each co
 | Module | Purpose | Key Classes/Functions | Lines |
 |--------|---------|----------------------|-------|
 | [`tool_mock.py`](tool_mock.py) | Function mocking & FastAPI MCP | `VerisSDK`, `@mock`, `@stub` | 327 |
-| [`braintrust_tracing.py`](braintrust_tracing.py) | Dual tracing instrumentation | `instrument()` | 283 |
 | [`utils.py`](utils.py) | Type utilities & JSON schema | `extract_json_schema()` | 272 |
 | [`logging.py`](logging.py) | Logging configuration | `setup_logging()` | 116 |
 | [`models.py`](models.py) | Data models | Type definitions | 12 |
@@ -40,12 +39,6 @@ This module contains the core implementation of the Veris AI Python SDK. Each co
 
 **Implementation**: [`tool_mock.py:250-300`](tool_mock.py)
 
-### Tracing Flow
-1. **Dual Setup**: Braintrust + OpenTelemetry instrumentation  
-2. **Session Tagging**: Bearer tokens → session IDs
-3. **Span Attribution**: All operations tagged with `veris.session_id`
-
-**Implementation**: [`braintrust_tracing.py:50-150`](braintrust_tracing.py)
 
 ## Configuration
 

veris_ai-1.6.0/src/veris_ai/__init__.py

@@ -0,0 +1,17 @@
+"""Veris AI Python SDK."""
+
+__version__ = "0.1.0"
+
+# Import lightweight modules that only use base dependencies
+from .jaeger_interface import JaegerClient
+from .models import ResponseExpectation
+from .tool_mock import veris
+from .observability import init_observability, instrument_fastapi_app
+
+__all__ = [
+    "veris",
+    "JaegerClient",
+    "ResponseExpectation",
+    "init_observability",
+    "instrument_fastapi_app",
+]

{veris_ai-1.4.0 → veris_ai-1.6.0}/src/veris_ai/jaeger_interface/__init__.py

@@ -34,6 +34,6 @@ are fully typed using *pydantic* models so that IDEs can provide proper
 autocomplete and type checking.
 """
 
-from .client import JaegerClient as JaegerClient  # noqa: F401
+from .client import JaegerClient
 
 __all__ = ["JaegerClient"]

{veris_ai-1.4.0 → veris_ai-1.6.0}/src/veris_ai/logging.py

@@ -21,6 +21,7 @@ async def log_tool_call_async(
     if not base_url:
         logger.warning("VERIS_ENDPOINT_URL not set, skipping tool call logging")
         return
+    base_url = base_url.rstrip("/")
 
     endpoint = f"{base_url}/api/v2/simulations/{session_id}/log_tool_call"
     payload = {
@@ -32,8 +33,17 @@ async def log_tool_call_async(
     timeout = float(os.getenv("VERIS_MOCK_TIMEOUT", "90.0"))
 
     try:
+        headers: dict[str, str] | None = None
+        try:
+            from opentelemetry.propagate import get_global_textmap
+
+            headers = {}
+            get_global_textmap().inject(headers)
+        except Exception:  # pragma: no cover - otel optional
+            headers = None
+
         async with httpx.AsyncClient(timeout=timeout) as client:
-            response = await client.post(endpoint, json=payload)
+            response = await client.post(endpoint, json=payload, headers=headers)
             response.raise_for_status()
             logger.debug(f"Tool call logged for {function_name}")
     except Exception as e:
@@ -62,15 +72,24 @@ def log_tool_call_sync(
     timeout = float(os.getenv("VERIS_MOCK_TIMEOUT", "90.0"))
 
     try:
+        headers: dict[str, str] | None = None
+        try:
+            from opentelemetry.propagate import get_global_textmap  # type: ignore[import-not-found]
+
+            headers = {}
+            get_global_textmap().inject(headers)
+        except Exception:  # pragma: no cover - otel optional
+            headers = None
+
         with httpx.Client(timeout=timeout) as client:
-            response = client.post(endpoint, json=payload)
+            response = client.post(endpoint, json=payload, headers=headers)
             response.raise_for_status()
             logger.debug(f"Tool call logged for {function_name}")
     except Exception as e:
         logger.warning(f"Failed to log tool call for {function_name}: {e}")
 
 
-async def log_tool_response_async(session_id: str, response: object) -> None:
+async def log_tool_response_async(session_id: str, response: Any) -> None:  # noqa: ANN401
     """Log tool response asynchronously to the VERIS logging endpoint."""
     base_url = os.getenv("VERIS_ENDPOINT_URL")
     if not base_url:
@@ -85,15 +104,24 @@ async def log_tool_response_async(session_id: str, response: object) -> None:
     timeout = float(os.getenv("VERIS_MOCK_TIMEOUT", "90.0"))
 
     try:
+        headers: dict[str, str] | None = None
+        try:
+            from opentelemetry.propagate import get_global_textmap  # type: ignore[import-not-found]
+
+            headers = {}
+            get_global_textmap().inject(headers)
+        except Exception:  # pragma: no cover - otel optional
+            headers = None
+
         async with httpx.AsyncClient(timeout=timeout) as client:
-            http_response = await client.post(endpoint, json=payload)
+            http_response = await client.post(endpoint, json=payload, headers=headers)
             http_response.raise_for_status()
             logger.debug("Tool response logged")
     except Exception as e:
         logger.warning(f"Failed to log tool response: {e}")
 
 
-def log_tool_response_sync(session_id: str, response: object) -> None:
+def log_tool_response_sync(session_id: str, response: Any) -> None:  # noqa: ANN401
     """Log tool response synchronously to the VERIS logging endpoint."""
     base_url = os.getenv("VERIS_ENDPOINT_URL")
     if not base_url:
@@ -108,8 +136,17 @@ def log_tool_response_sync(session_id: str, response: object) -> None:
     timeout = float(os.getenv("VERIS_MOCK_TIMEOUT", "90.0"))
 
     try:
+        headers: dict[str, str] | None = None
+        try:
+            from opentelemetry.propagate import get_global_textmap  # type: ignore[import-not-found]
+
+            headers = {}
+            get_global_textmap().inject(headers)
+        except Exception:  # pragma: no cover - otel optional
+            headers = None
+
         with httpx.Client(timeout=timeout) as client:
-            http_response = client.post(endpoint, json=payload)
+            http_response = client.post(endpoint, json=payload, headers=headers)
             http_response.raise_for_status()
             logger.debug("Tool response logged")
     except Exception as e: