blocker-detector

Blocker Detector

Detect recurring blockers and their resolutions from conversation history.

Purpose

Analyze conversations to extract:

• Blocking issues encountered
• Solutions that worked
• Recurring problem patterns
• Prevention strategies

Detection Keywords

Problem Indicators

PROBLEM_KEYWORDS = [
    "doesn't work", "not working", "failed", "error", "bug",
    "issue", "problem", "broken", "crash", "exception",
    "wrong", "incorrect", "unexpected", "confused", "stuck",
    "can't", "cannot", "unable", "impossible", "blocked"
]

Solution Indicators

SOLUTION_KEYWORDS = [
    "fixed", "solved", "solution", "works now", "working now",
    "the fix", "resolved", "figured out", "found the issue",
    "the answer", "turns out", "the trick", "workaround",
    "finally", "success", "got it working"
]

Analysis Method

1. Problem Detection

for msg in user_messages:
    if any(kw in msg.lower() for kw in PROBLEM_KEYWORDS):
        problems.append({
            'content': msg,
            'session_id': session_id,
            'timestamp': timestamp
        })

2. Solution Pairing

# Look for solutions in subsequent messages
for i, problem in enumerate(problems):
    for j in range(i+1, min(i+10, len(messages))):
        if any(kw in messages[j].lower() for kw in SOLUTION_KEYWORDS):
            problem['solution'] = messages[j]
            break

3. Categorization

CATEGORIES = {
    "remote_execution": ["remote", "psexec", "ssh", "session_id"],
    "hdf_files": ["hdf", "h5py", "results"],
    "geometry": ["geometry", "cross section", "mesh"],
    "documentation": ["documentation", "mkdocs", "notebook"],
    "git": ["git", "commit", "merge", "branch"],
    "imports": ["import", "module", "package"]
}

def categorize(text):
    for category, keywords in CATEGORIES.items():
        if any(kw in text.lower() for kw in keywords):
            return category
    return "general"