log_stream.py

"""
Thread-safe log queue + custom logging handler.
Log records are pushed into a queue; the SSE endpoint drains it to the browser.
"""
import logging
import queue
import time
from datetime import datetime

_log_queue: queue.Queue = queue.Queue(maxsize=500)

LEVEL_EMOJI = {
    "DEBUG":    "🔍",
    "INFO":     "ℹ️",
    "WARNING":  "⚠️",
    "ERROR":    "❌",
    "CRITICAL": "🚨",
}

STEP_MARKERS = {
    "Starting daily recommendation pipeline": ("🚀", "Pipeline started"),
    "Fetching Scholar profile":               ("👤", "Loading your Scholar profile"),
    "Using cached Scholar profile":           ("👤", "Scholar profile loaded from cache"),
    "Fetching arxiv papers":                  ("📡", "Fetching latest arXiv papers"),
    "Got first page":                         ("📄", "arXiv papers received"),
    "Topic signals":                          ("🗳️",  "Loading vote signals"),
    "ML ranker":                              ("🤖", "Training ML preference model"),
    "TF-IDF model":                           ("🤖", "TF-IDF model ready"),
    "Computing reputation":                   ("🏛️",  "Scoring institution reputation"),
    "HTTP Request: POST https://api.anthropic": ("✨", "Claude is ranking papers…"),
    "Summarized:":                            ("📝", "Generating summary"),
    "Saved":                                  ("💾", "Recommendations saved"),
    "Pipeline complete":                      ("✅", "Pipeline complete!"),
}


class QueueHandler(logging.Handler):
    def emit(self, record: logging.LogRecord):
        msg = self.format(record)
        level = record.levelname
        emoji = LEVEL_EMOJI.get(level, "•")

        # Detect named step
        step_label = None
        for marker, (step_emoji, label) in STEP_MARKERS.items():
            if marker in msg:
                emoji = step_emoji
                step_label = label
                break

        entry = {
            "ts": datetime.now().strftime("%H:%M:%S"),
            "level": level,
            "emoji": emoji,
            "step": step_label,
            "msg": record.getMessage(),
        }
        try:
            _log_queue.put_nowait(entry)
        except queue.Full:
            pass  # drop oldest would be nicer but this is fine


def get_queue() -> queue.Queue:
    return _log_queue


def install_handler():
    """Attach QueueHandler to the root logger."""
    handler = QueueHandler()
    handler.setLevel(logging.DEBUG)
    formatter = logging.Formatter("%(name)s: %(message)s")
    handler.setFormatter(formatter)
    logging.getLogger().addHandler(handler)