Capture Thread and Process Context in Spans

[apiad]

Description:

To visualize "where" code is executing in the proposed TUI (ferret watch), we need to know which thread and process owns each span. Currently, SpanModel only captures timing and hierarchy data. Without this, the live monitor cannot group active spans by thread (e.g., "Thread-1 is executing fetch_db", "MainProcess is executing setup") or distinguish between multiple worker processes running in parallel.

Proposed Solution:

1. Schema Update: Add thread_id, thread_name, and process_id fields to the SpanModel.
2. Capture Logic: During Span initialization in the Profiler, automatically capture these values using Python's standard os and threading libraries.

Tasks:

• Update Model (ferret/models.py):
  • Add fields to SpanModel:
    • process_id: int
    • thread_id: int
    • thread_name: str (Optional, but highly recommended for TUI readability)
• Update Core (ferret/core.py):
  • In Span.__init__, capture the context:

    import os
    import threading
    
    # ... inside __init__
    self.model.process_id = os.getpid()
    self.model.thread_id = threading.get_ident()
    self.model.thread_name = threading.current_thread().name

• TUI Integration Prep:
  • Ensure these fields are indexed or easily accessible so the future ferret watch app can group the "Live Stack" by process_id -> thread_id.

Acceptance Criteria:

• Inspecting the ferret.db (via ferret analyze or raw check) shows that every span now contains valid PID and Thread ID information.
• Spans generated from asyncio tasks correctly reflect the thread running the event loop (usually MainThread).
• Spans generated from concurrent.futures.ThreadPoolExecutor show distinct thread_ids.