from __future__ import annotations

import re
from urllib.parse import parse_qs, urlparse


REASON_CLINIC_DATA_NOT_READY = 'clinic_data_not_ready'
REASON_GRAFANA_RELATIVE_WINDOW = 'grafana_relative_time_window'
REASON_PER_STORE_EVIDENCE_MISSING = 'per_store_evidence_missing'
REASON_PER_HOP_EVIDENCE_MISSING = 'per_hop_evidence_missing'
REASON_LOWER_LAYER_EVIDENCE_MISSING = 'lower_layer_evidence_missing'


def grafana_link_has_relative_window(link: str) -> bool:
    parsed = urlparse(str(link or '').strip())
    params = parse_qs(parsed.query)
    for key in ('from', 'to'):
        for value in params.get(key, []):
            normalized = str(value or '').strip().lower()
            if normalized.startswith('now'):
                return True
    return False


def analyze_missing_evidence_reasons(context: dict, clinic_results: list[dict]) -> list[str]:
    reasons = []
    text = '\n'.join(
        str(context.get(key) or '')
        for key in ('title', 'problem_statement', 'phenomenon', 'routing_context_text')
    )
    lowered = text.lower()
    successful_clinic = [item for item in clinic_results or [] if not item.get('error')]

    if any((item.get('data_status') or {}) and not bool((item.get('data_status') or {}).get('ready')) for item in clinic_results or []):
        reasons.append(REASON_CLINIC_DATA_NOT_READY)

    grafana_refs = context.get('grafana_references') or []
    if grafana_refs and any(grafana_link_has_relative_window(item.get('link') or '') for item in grafana_refs):
        reasons.append(REASON_GRAFANA_RELATIVE_WINDOW)

    if not successful_clinic and 'tikv' in lowered:
        reasons.append(REASON_PER_STORE_EVIDENCE_MISSING)

    if re.search(
        r'tikvrpc|batch receive|wait connection establish|rpc layer latency|unreachable|'
        r'network jitter|network latency|network interruption|network saturation|'
        r'带宽打满|网络抖动|网络异常|网络问题',
        text,
        re.I,
    ):
        reasons.append(REASON_PER_HOP_EVIDENCE_MISSING)

    if re.search(
        r'serverisbusy|server is busy|deadline exceeded|flow control|throttle duration|'
        r'discard ratio|scheduler pending write|pending compaction|l0 file|memtable|'
        r'snapshot|scatter|rebalance|delete-range|io jitter|disk latency|ebs|rocksdb',
        text,
        re.I,
    ):
        reasons.append(REASON_LOWER_LAYER_EVIDENCE_MISSING)

    deduped = []
    seen = set()
    for reason in reasons:
        key = str(reason or '').strip()
        if not key or key in seen:
            continue
        seen.add(key)
        deduped.append(key)
    return deduped