from __future__ import annotations

import json

from investibook.core.context_debug import build_context_debug_report


def _category(report: dict, key: str) -> dict:
    for item in report["categories"]:
        if item["key"] == key:
            return item
    raise AssertionError(f"missing category {key}")


def test_context_debug_report_splits_context_sources(tmp_path, monkeypatch):
    workspace_dir = tmp_path / "workspaces" / "ws-test"
    workspace_dir.mkdir(parents=True)

    prompt_dir = tmp_path / "prompts"
    prompt_dir.mkdir()
    instruction_file = prompt_dir / "legacy_instruction.md"
    instruction_file.write_text("static instruction text", encoding="utf-8")
    monkeypatch.setattr("investibook.config.AGENTKIT_PROMPTS_DIR", prompt_dir)

    opencode_dir = workspace_dir / ".opencode"
    opencode_dir.mkdir()
    (opencode_dir / "opencode.json").write_text(
        json.dumps({"instructions": [str(instruction_file)]}),
        encoding="utf-8",
    )

    enabled = opencode_dir / "skills" / "builtin" / "demo"
    enabled.mkdir(parents=True)
    (enabled / "SKILL.md").write_text("skill instructions", encoding="utf-8")
    knowledge = enabled / "knowledge"
    knowledge.mkdir()
    (knowledge / "pitfalls.md").write_text("knowledge text", encoding="utf-8")

    all_skills = opencode_dir / ".all-skills" / "huge"
    all_skills.mkdir(parents=True)
    (all_skills / "SKILL.md").write_text("x" * 10000, encoding="utf-8")

    issue_dir = workspace_dir / "issues" / "O11Y-1"
    issue_dir.mkdir(parents=True)
    (issue_dir / "cluster_meta.json").write_text(
        json.dumps(
            {
                "cluster_id": "c1",
                "org_id": "o1",
                "tidb_version": "v8.5.0",
                "alert_time": "2026-05-09T07:00:00+00:00",
            }
        ),
        encoding="utf-8",
    )

    evidence_dir = workspace_dir / "evidences"
    evidence_dir.mkdir()
    (evidence_dir / "evidence_summary_for_agent.txt").write_text(
        "important evidence summary",
        encoding="utf-8",
    )

    messages = [
        {"id": "u1", "role": "user", "content": "diagnose", "timestamp": "t1"},
        {
            "id": "issue-info-O11Y-1",
            "role": "system",
            "content": "O11Y-1 cluster c1 alert metadata",
            "timestamp": "t1",
            "issueInfo": {"id": "O11Y-1", "cluster_id": "c1"},
        },
        {
            "id": "a1",
            "role": "assistant",
            "content": "answer",
            "timestamp": "t2",
            "parts": [
                {
                    "type": "tool",
                    "tool": "skill",
                    "state": {
                        "title": "Loaded skill: demo",
                        "input": {"name": "demo"},
                        "output": "skill instructions",
                    },
                },
                {
                    "type": "tool",
                    "tool": "read",
                    "state": {
                        "input": {"filePath": str(knowledge / "pitfalls.md")},
                        "output": "knowledge text",
                    },
                },
                {
                    "type": "tool",
                    "tool": "Read",
                    "state": {
                        "input": {"path": "evidences/evidence_summary_for_agent.txt"},
                        "output": "important evidence summary",
                    },
                },
                {
                    "type": "reasoning",
                    "text": "visible thinking text",
                }
            ],
        },
    ]
    usage_records = [
        {
            "session_id": "evidence-postpass:a1",
            "channel": "evidence-postpass",
            "model_tier": "evidence-postpass",
            "context_used": 0,
            "context_size": 0,
        },
        {
            "session_id": "ses-real",
            "channel": "web",
            "model_tier": "standard",
            "context_used": 30000,
            "context_size": 100000,
            "created_at": "2026-05-09T07:17:21+00:00",
            "reasoning_tokens": 1234,
            "cache_read_tokens": 29900,
        },
    ]

    report = build_context_debug_report(
        "ws-test",
        workspace_dir=workspace_dir,
        usage_records=usage_records,
        messages=messages,
    )

    assert report["actual_context"]["context_used"] == 30000
    assert report["actual_context"]["context_size"] == 100000
    assert report["actual_context"]["percent"] == 30.0
    assert report["actual_context"]["reasoning_tokens"] == 1234

    assert _category(report, "agent_runtime_instructions")["included_in_context_estimate"] is True
    assert _category(report, "skills")["included_in_context_estimate"] is True
    assert _category(report, "evidence")["included_in_context_estimate"] is True
    assert _category(report, "cluster_alert_meta")["included_in_context_estimate"] is True
    assert _category(report, "conversation")["included_in_context_estimate"] is True
    assert "assistant_thinking" not in {row["key"] for row in report["categories"]}

    evidence_files = _category(report, "evidence_available")
    assert evidence_files["scope"] == "available-file"
    assert evidence_files["scaled_context_tokens"] is None

    evidence_examples = _category(report, "evidence")["examples"]
    assert evidence_examples[0]["label"] == "Read evidence summary: evidence_summary_for_agent.txt"
    assert evidence_examples[0]["detail"] == "assistant message part #3"
    assert evidence_examples[0]["source_ref"] == "messages_v2/a1/parts/2"
    assert evidence_examples[0]["percent_of_used_context"] is not None
    assert evidence_examples[0]["scaled_context_tokens"] is not None

    skill_examples = _category(report, "skills")["examples"]
    skill_labels = {example["label"] for example in skill_examples}
    assert "Loaded skill: demo" in skill_labels
    assert "Read skill knowledge: demo/knowledge/pitfalls.md" in skill_labels
    assert not any(example["label"] == "SKILL.md" for example in skill_examples)

    meta_examples = _category(report, "cluster_alert_meta")["examples"]
    assert any(example["label"] == "Jira issue metadata: O11Y-1" for example in meta_examples)

    knowledge_files = _category(report, "skills_available")
    assert knowledge_files["scope"] == "available-file"
    assert knowledge_files["scaled_context_tokens"] is None
    assert any(example["label"] == "SKILL.md" for example in knowledge_files["examples"])

    cluster_files = _category(report, "cluster_alert_meta_available")
    assert cluster_files["scope"] == "available-file"
    assert cluster_files["scaled_context_tokens"] is None

    category_keys = {item["key"] for item in report["categories"]}
    assert "skill_support_files_available" not in category_keys
    assert "chat_tool_evidence_payloads" not in category_keys
    assert not any(
        "huge" in example["path"]
        for item in report["categories"]
        for example in item["examples"]
    )