File size: 10,167 Bytes
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
a8124a8
 
 
 
 
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a8124a8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a8124a8
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31e3087
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31e3087
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31e3087
 
 
 
 
 
 
 
60a2fc8
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
#!/usr/bin/env python3
"""Build the interactive research-roadmap data contract for the public site."""

from __future__ import annotations

import json
from datetime import datetime, timezone
from pathlib import Path
from typing import Any


ROOT = Path(__file__).resolve().parents[1]
DOCS_DATA = ROOT / "docs" / "data"
RESULTS = ROOT / "results" / "episode_task_suite"
GITHUB_BLOB = "https://github.com/ChaoYue0307/ropedia-xperience-10m-task-suite/blob/main"


def repo_link(path: str) -> str:
    return f"{GITHUB_BLOB}/{path}"


def load_json(path: Path) -> Any:
    return json.loads(path.read_text(encoding="utf-8"))


def rounded_metric(value: Any) -> float | None:
    if value is None:
        return None
    try:
        return round(float(value), 4)
    except (TypeError, ValueError):
        return None


def metric_summary(metric: dict[str, Any] | None) -> dict[str, Any]:
    if not metric:
        return {}
    return {
        "key": metric.get("key"),
        "name": metric.get("name"),
        "direction": metric.get("direction"),
        "minimal": rounded_metric(metric.get("minimal")),
        "neural_mlp": rounded_metric(metric.get("neural_mlp")),
        "better_baseline": metric.get("better_baseline"),
    }


def task_evidence_links(task_id: str) -> list[dict[str, str]]:
    candidates = [
        ("Minimal metrics", f"results/episode_task_suite/{task_id}/metrics.json"),
        ("Neural metrics", f"results/episode_task_suite/neural_mlp/{task_id}/metrics.json"),
        ("Minimal predictions", f"results/episode_task_suite/{task_id}/predictions.csv"),
        ("Neural predictions", f"results/episode_task_suite/neural_mlp/{task_id}/predictions.csv"),
        ("Confusion matrix", f"results/episode_task_suite/{task_id}/confusion_matrix.csv"),
        ("Neural confusion matrix", f"results/episode_task_suite/neural_mlp/{task_id}/confusion_matrix.csv"),
    ]
    links = [
        {"label": "Task walkthrough", "href": "data/task_walkthroughs.json"},
        {"label": "Single-episode explorer", "href": "single_episode_explorer.html"},
    ]
    for label, relative_path in candidates:
        if (ROOT / relative_path).exists():
            links.append({"label": label, "href": repo_link(relative_path)})
    return links


def task_payload(
    task_id: str,
    direction_task: dict[str, Any],
    walkthrough: dict[str, Any],
) -> dict[str, Any]:
    metric = direction_task.get("metric") or walkthrough.get("metric") or {}
    return {
        "id": task_id,
        "display_name": walkthrough.get("display_name") or direction_task.get("name") or task_id,
        "research_name": walkthrough.get("research_name") or direction_task.get("name") or task_id,
        "family": direction_task.get("family") or walkthrough.get("task_family"),
        "architecture_family": walkthrough.get("architecture_family"),
        "primary_direction": direction_task.get("primary_direction"),
        "direction_roles": direction_task.get("direction_roles", {}),
        "modalities": walkthrough.get("modalities", []),
        "case_study": walkthrough.get("case_study"),
        "input": walkthrough.get("input"),
        "input_short": walkthrough.get("input_short"),
        "process_short": walkthrough.get("process_short"),
        "output_short": walkthrough.get("output_short"),
        "module_summary": walkthrough.get("module_summary"),
        "current_limit": direction_task.get("current_limit") or walkthrough.get("failure_mode"),
        "why": direction_task.get("why"),
        "metric": metric_summary(metric),
        "evidence_links": task_evidence_links(task_id),
    }


def phase_payload(phases: list[dict[str, Any]]) -> list[dict[str, Any]]:
    stage_map = {
        "implemented": "now",
        "active": "scale_up",
        "next": "omni",
        "planned": "future",
    }
    return [
        {
            "id": phase.get("id"),
            "name": phase.get("name"),
            "status": phase.get("status"),
            "stage": stage_map.get(str(phase.get("status", "")).lower(), "future"),
            "entry_condition": phase.get("entry_condition"),
            "deliverables": phase.get("deliverables", []),
            "completion_evidence": phase.get("completion_evidence", []),
            "reader_takeaway": phase.get("reader_takeaway"),
        }
        for phase in phases
    ]


def main() -> int:
    directions_doc = load_json(DOCS_DATA / "research_directions.json")
    walkthroughs = load_json(DOCS_DATA / "task_walkthroughs.json")
    roadmap = load_json(DOCS_DATA / "research_roadmap.json")
    foundation_plan = load_json(DOCS_DATA / "foundation_model_plan.json")
    summary_metrics = load_json(DOCS_DATA / "summary_metrics.json")
    episode_summary = load_json(RESULTS / "summary_report.json")
    feature_manifest = load_json(RESULTS / "feature_manifest.json")
    extension_doc = load_json(DOCS_DATA / "research_direction_extensions.json")

    tasks: dict[str, dict[str, Any]] = {}
    for task_id, direction_task in directions_doc.get("tasks", {}).items():
        tasks[task_id] = task_payload(
            task_id,
            direction_task,
            walkthroughs.get("tasks", {}).get(task_id, {}),
        )

    directions = []
    for code, direction in directions_doc.get("directions", {}).items():
        linked_tasks = [tasks[task_id] for task_id in direction.get("tasks", []) if task_id in tasks]
        extension_tasks = [
            {
                "id": task_id,
                "name": spec.get("name"),
                "family": spec.get("family"),
                "metric_name": spec.get("metric_name"),
                "current_limit": spec.get("current_limit"),
            }
            for task_id, spec in extension_doc.get("task_specs", {}).items()
            if spec.get("direction") == code
        ]
        directions.append(
            {
                "code": code,
                "id": direction.get("id"),
                "name": direction.get("name"),
                "focus": direction.get("focus"),
                "preferred_background": direction.get("preferred_background"),
                "current_status": direction.get("current_status"),
                "current_readout": direction.get("current_readout"),
                "next_steps": direction.get("next_steps", []),
                "counts": direction.get("counts", {}),
                "task_ids": direction.get("tasks", []),
                "tasks": linked_tasks,
                "extension_tasks": extension_tasks,
            }
        )

    omni = summary_metrics.get("omni_relay", {})
    payload = {
        "title": "Interactive Research Roadmap",
        "generated_at_utc": datetime.now(timezone.utc).isoformat(timespec="seconds"),
        "source_files": [
            "docs/data/research_directions.json",
            "docs/data/task_walkthroughs.json",
            "docs/data/research_roadmap.json",
            "docs/data/foundation_model_plan.json",
            "docs/data/summary_metrics.json",
            "docs/data/research_direction_extensions.json",
            "results/episode_task_suite/summary_report.json",
            "results/episode_task_suite/feature_manifest.json",
        ],
        "scope": {
            "sample_episode_count": walkthroughs.get("scope", {}).get("episode_count", 1),
            "num_frames": episode_summary.get("num_frames"),
            "num_windows": episode_summary.get("num_windows"),
            "feature_dim": episode_summary.get("feature_dim"),
            "window_frames": episode_summary.get("window_frames"),
            "stride_frames": episode_summary.get("stride_frames"),
            "feature_blocks": len(feature_manifest),
            "warning": walkthroughs.get("scope", {}).get("warning"),
        },
        "baseline_summary": {
            "task_count": len(tasks),
            "baseline_heads": "minimal and neural MLP heads",
            "split": "chronological single-episode split for public-sample diagnostics",
            "current_use": "task design, data-contract validation, case studies, and baseline comparison",
        },
        "scale_up": {
            "target_episodes": omni.get("target_episodes"),
            "candidate_scan_top_level_sessions": omni.get("candidate_scan_top_level_sessions"),
            "valid_candidates": omni.get("valid_candidates"),
            "estimated_bytes": omni.get("estimated_bytes"),
            "status": omni.get("status"),
            "access_status": omni.get("access_status"),
            "exclude": omni.get("exclude", []),
            "selection_strategy": omni.get("selection_strategy"),
        },
        "omni_plan": {
            "backbone": "Qwen/Qwen3-Omni-30B-A3B-Instruct",
            "adapter": "LoRA rank 16, alpha 32, dropout 0.05",
            "first_pilot": "32 held-out-episode pilot after valid episodes are staged",
            "training_unit": "episode-level split, window-level supervised examples",
            "evaluation": [
                "JSON validity",
                "action macro-F1",
                "subtask accuracy",
                "transition accuracy",
                "next-action accuracy",
                "contact accuracy",
                "object micro-F1",
                "held-out episode count",
            ],
        },
        "foundation_model_plan": {
            "status": foundation_plan.get("status"),
            "decision": foundation_plan.get("decision", {}),
            "model_families": foundation_plan.get("model_families", []),
            "execution_order": foundation_plan.get("execution_order", []),
            "evaluation_additions": foundation_plan.get("evaluation_additions", []),
            "source_links": foundation_plan.get("source_links", []),
        },
        "phases": phase_payload(roadmap.get("phases", [])),
        "directions": directions,
        "tasks": list(tasks.values()),
    }

    out_path = DOCS_DATA / "research_roadmap_interactive.json"
    out_path.write_text(json.dumps(payload, indent=2, sort_keys=True) + "\n", encoding="utf-8")
    print(f"Wrote {out_path}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())