coyote/docs/testing/plans/07-input-construction.md

Test Plan: Input Construction

Feature description

Input encapsulates a single chat turn's data: text, files, role, model, session context, RAG embeddings, and function declarations. It's constructed at the start of each turn and captures all needed state from RequestContext.

Behaviors to test

Input::from_str

• Creates Input from text string
• Captures role via resolve_role
• Captures session from ctx
• Captures rag from ctx (requires RAG setup)
• Captures functions via select_functions (tested separately)
• Captures stream_enabled from AppConfig
• app_config field set from ctx.app.config
• Empty text → is_empty() returns true

Input::from_files

• Loads file contents (async + filesystem)
• Supports multiple files (async + filesystem)
• Supports directories (recursive) (async + filesystem)
• Supports URLs (fetches content) (async + network)
• Supports loader syntax (e.g., jina:url) (async + loader)
• Last message carry-over (%% syntax) (via resolve_paths)
• Combines file content with text (async)
• document_loaders from AppConfig used (async)

resolve_role

• Returns provided role if given
• Extracts role from agent if agent active (requires agent init)
• Extracts role from session if session has role
• Returns default model-based role otherwise
• with_session flag set correctly
• with_agent flag set correctly

Input methods

• stream() returns stream_enabled && !model.no_stream() (requires Model with no_stream)
• create_client() uses app_config to init client (requires client config)
• prepare_completion_data() uses captured functions (requires Model)
• build_messages() uses captured session (requires Message setup)
• echo_messages() uses captured session (requires Message setup)
• set_regenerate(role) refreshes role
• use_embeddings() searches RAG if present (requires RAG)
• merge_tool_results() creates continuation input (requires ToolResult)

Context switching scenarios

• Input with agent → agent functions selected (requires agent init)
• Input with MCP → MCP meta functions in declarations (via select_functions tests)
• Input with RAG → embeddings included after use_embeddings (requires RAG)
• Input without session → no session messages in build_messages (via session() test)

Additional behaviors tested (not in original plan)

• resolve_role: explicit role overrides session flag
• resolve_paths: empty input
• resolve_paths: URL detection (https://)
• resolve_paths: external command detection (backtick syntax)
• resolve_paths: rejects URL with glob suffix
• resolve_paths: mixed inputs (%%, URL, external cmd)
• Input::set_text changes text
• Input::patched_text overrides text()
• Input::clear_patch restores original
• Input::set_continue_output accumulates
• Input::summary truncates long text with ...
• Input::summary preserves short text
• Input::raw() with no files
• Input::render() with no medias
• Input::session() returns None when with_session=false
• Input::session() returns Some when with_session=true
• is_image recognizes png/jpeg/jpg/webp/gif
• is_image rejects non-image extensions
• resolve_data_url returns path for known hash
• resolve_data_url returns original for non-data URL
• select_functions: None when no tools enabled
• select_functions: None when function_calling disabled
• select_functions: "all" returns all non-MCP
• select_functions: comma-separated filters
• select_enabled_mcp_servers: empty when MCP disabled
• select_enabled_mcp_servers: "all" returns all MCP functions
• select_enabled_mcp_servers: comma filters by server name

Old code reference

• src/config/input.rs — Input struct, from_str, from_files
• src/config/mod.rs — select_functions, extract_role