library · inferred from evidence
llama-index-core
A library used alongside ariz-phoenix in the integration.
- [Bug]: WorkflowRuntimeError: Error in step 'run_agent_step': cannot access local variable 'r' where it is not associated
LlamaIndex: Collaborator logan-markewich reproduced the multi-agent ReAct bug: the workflow crashed when the LLM streamed zero data in run_agent_step. Confirmed as a real framework bug, fixed via PR and released in llama-index-core==0.12.14
- Role 'tool' must be a response to a preceding message with 'tool_calls'
LlamaIndex: Collaborator logan-markewich: this was a known bug in the chat-store memory handling - the memory buffer dropped the assistant/tool message pairs, so a 'tool' role message was sent to OpenAI without a preceding 'tool_calls' as
- [Bug]: cannot import name 'LLM' from 'llama_index.core.llms'
LlamaIndex: Collaborator logan-markewich: the v0.10 split into namespaced packages (llama-index-core + integrations) means any remnants of an old 0.9.x install - globally, outside the venv, or pulled in by another dependency - break imp
- LlamaParse `output_tables_as_HTML` has no effect on `items[].md`
How to fix llamaparse: llamaparse `output_tables_as_html` has no effect on `items[].md`.
- LlamaParse `output_tables_as_HTML` has no effect on `items[].md`
Expected behavior, not a bug. When you use the structured `Document`/`Page`/`Item` result, `item.md` is always converted back to markdown , even when the raw LlamaParse output contains HTML table tags , to stay consistent with the llama-ind
- Phoenix protobuf error on import: Descriptors cannot be created directly
The real problem is pip backtracking to an old `opentelemetry-proto` (1.11.1) which only requires protobuf>=3.13, incompatible with protobuf v5. Fix: constrain `opentelemetry-exporter-otlp >=1.21.0` (or update all of `opentelemetry-proto`,