[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-tao-route-visual-changenet-samples":3,"mdc--4ta7nl-key":31,"related-org-nvidia-tao-route-visual-changenet-samples":1408,"related-repo-nvidia-tao-route-visual-changenet-samples":1566},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":20,"repoUrl":21,"updatedAt":22,"license":23,"forks":24,"topics":25,"repo":26,"sourceUrl":29,"mdContent":30},"tao-route-visual-changenet-samples","route Visual ChangeNet samples for augmentation","Routes the weakest VCN samples (output of `tao-analyze-gaps-visual-changenet`) into per-augmentation-module subsets based on each module's label eligibility. Use when the user asks to \"route VCN gap samples\", \"split AOI gaps for k-NN mining and AnomalyGen\", or prepare the immediate next step after DEFT gap analysis in a VCN AOI SDA iteration.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19],{"name":13,"slug":14,"type":15},"Deep Learning","deep-learning","tag",{"name":17,"slug":18,"type":15},"Automation","automation",{"name":9,"slug":8,"type":15},2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:30:45.320502","Apache-2.0",281,[],{"repoUrl":21,"stars":20,"forks":24,"topics":27,"description":28},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Ftao-route-visual-changenet-samples","---\nname: tao-route-visual-changenet-samples\ndescription: Routes the weakest VCN samples (output of `tao-analyze-gaps-visual-changenet`) into per-augmentation-module\n  subsets based on each module's label eligibility. Use when the user asks to \"route VCN gap samples\", \"split AOI gaps for\n  k-NN mining and AnomalyGen\", or prepare the immediate next step after DEFT gap analysis in a VCN AOI SDA iteration.\nlicense: Apache-2.0\ncompatibility: Standalone — no external runtime requirements.\nmetadata:\n  author: NVIDIA Corporation\n  version: \"0.1.0\"\nallowed-tools: Read Bash\ntags:\n- data\n- routing\n- vcn\n- aoi\n- sda\n---\n\n# TAO VCN Sample Routing Skill\n\nYou are the dispatcher between gap analysis and the augmentation modules in a VCN AOI SDA pipeline. Each augmentation module can only act on labels it knows how to handle:\n\n- **k-NN Mining** can only mine real-image neighbors for labels that already exist in the **source pool CSV**. There is no point looking for `SHIFT` neighbors if the pool has no `SHIFT` rows.\n- **AnomalyGen** (Cosmos SDG) can only generate synthetic anomalies for the classes its inference pipeline supports: `PASS`, `EXCESS_SOLDER`, `MISSING`, `BRIDGE`. A weak sample with a label outside this set is unroutable to AnomalyGen.\n\nThis skill runs **once per SDA iteration immediately after gap analysis**. It splits the gap-analysis parquet into one filtered parquet per module so each module operates on its own eligible subset, and it writes a human-readable summary of the per-label routing decisions.\n\nThe work is intentionally trivial: read a parquet, do two `.isin(...)` filters, write two parquets, write one summary. The skill exists to make those decisions auditable — every label must show up in the summary with a yes\u002Fno verdict for each module so a downstream reviewer can spot when a label is silently dropped because no module accepted it.\n\n---\n\n## Inputs\n\n1. **`gaps_parquet`** — the gap-analysis output (typically `\u003Cexp_dir>\u002Frca_results\u002F\u003Ctimestamp>\u002Fgaps.parquet` from `tao-analyze-gaps-visual-changenet`). Required columns: `filepath`, `label`. Other columns (`siamese_score`, `weakness`) are preserved verbatim.\n2. **`source_pool_csv`** — VCN-format mining source pool CSV with a `label` column. Empty string or non-existent path is allowed; the mining subset will simply be empty in that case.\n3. **Output directory** — where the two routed parquets, the summary, and the report are written. Default: a timestamped folder under the gap-analysis result directory: `\u003Crca_result_dir>\u002Frouting_results\u002F\u003Ctimestamp>\u002F`.\n4. **`anomalygen_supported_labels`** *(optional)* — override the default AnomalyGen-eligible label set. Default: `{\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}`. **Warning:** This must stay in sync with `ANOMALYGEN_SUPPORTED_LABELS` in `mdo-kratos-workflows\u002Fpipelines\u002Fsda\u002Frouting.py` and the AnomalyGen integration's actual generator coverage. Adding a new defect class to AnomalyGen means adding it here too.\n\n---\n\n## Method\n\nThe whole skill is two `.isin(...)` masks against the uppercased label column.\n\n### Step 1 — Load and uppercase\n\n```python\ndf = pd.read_parquet(gaps_parquet)\nlabels_upper = df[\"label\"].astype(str).str.upper()\n```\n\nThe match is **case-insensitive** for both module checks. The original `label` column is preserved unchanged in the output parquets — only the comparison key is uppercased.\n\n### Step 2 — Mining subset\n\n```python\nif source_pool_csv and os.path.isfile(source_pool_csv):\n    pool_df = pd.read_csv(source_pool_csv)\n    pool_labels = {str(l).upper() for l in pool_df[\"label\"].unique()}\n    mn_mask = labels_upper.isin(pool_labels)\n    mn_df = df[mn_mask]\nelse:\n    pool_missing = True\n    pool_labels = set()\n    mn_df = df.iloc[0:0]   # empty, but with the same schema\nmn_df.to_parquet(mining_gaps_parquet, index=False)\n```\n\nIf the pool CSV is missing or empty, the mining subset is an empty DataFrame **with the same columns as the input** so downstream readers don't crash on schema mismatch. Flag this case in the summary.\n\n### Step 3 — AnomalyGen subset\n\n```python\nANOMALYGEN_SUPPORTED = {\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}\nag_mask = labels_upper.isin(ANOMALYGEN_SUPPORTED)\nag_df = df[ag_mask]\nag_df.to_parquet(anomalygen_gaps_parquet, index=False)\n```\n\nRows whose label is in the AnomalyGen-supported set are written verbatim to `anomalygen_gaps.parquet`. The schema matches the input parquet exactly — downstream AnomalyGen (Cosmos SDG) needs no other changes.\n\n### Step 4 — Per-label routing breakdown\n\nFor every distinct label in the input gaps parquet (uppercased), record:\n- `count` — how many rows have this label\n- `mining` — yes if the label is in `pool_labels`, otherwise no\n- `anomalygen` — yes if the label is in `ANOMALYGEN_SUPPORTED`, otherwise no\n\nA label can route to **both** modules (e.g. PASS rows route to AnomalyGen, and if the source pool also contains PASS rows they route to Mining too). A label can also route to **none** — flag those, since they are silently dropped and may signal a configuration mismatch.\n\nWrite the breakdown to `routing_summary.txt`. The format mirrors the reference component exactly:\n\n```\nWeak-sample routing summary\nTotal weak samples: \u003CN>\nMining subset:      \u003CN_mn> -> \u003Cmining_gaps_parquet>\nAnomalyGen subset:  \u003CN_ag> -> \u003Canomalygen_gaps_parquet>\n\n[If pool missing:]\nNo source pool CSV at '\u003Cpath>'; mining subset is empty.\n\nPer-label breakdown (count, mining, anomalygen):\n  PASS: 50 (mining=yes, anomalygen=yes)\n  MISSING: 32 (mining=no, anomalygen=yes)\n  SHIFT: 14 (mining=yes, anomalygen=no)\n  EXCESS_SOLDER: 9 (mining=yes, anomalygen=yes)\n  ...\n```\n\n### Step 5 — Sanity checks\n\nAfter both subsets are written, verify:\n- The sum of subset sizes is *not* required to equal `len(df)` — overlap is allowed (a label can route to both modules). What matters is that **every input row appears in at least one subset, OR appears in the \"none\" list with an explicit reason**.\n- If `len(mn_df) == 0` and `len(ag_df) == 0`, something is wrong — flag prominently in the report.\n- If an entire label group routes to no module, the `Recommended Actions` section must call this out so the user can either seed the source pool with that label or extend AnomalyGen's supported set.\n\n---\n\n## Reference Python Recipe\n\nThis is the exact computation, lifted from `mdo-kratos-workflows\u002Fpipelines\u002Fsda\u002Frouting.py`. Run as a single Python script via Bash; it produces every artifact except the report.\n\n```python\nimport os\nimport pandas as pd\n\nANOMALYGEN_SUPPORTED = {\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}\n\ndf = pd.read_parquet(gaps_parquet)\nlabels_upper = df[\"label\"].astype(str).str.upper()\n\n# Mining subset\npool_missing = False\nif source_pool_csv and os.path.isfile(source_pool_csv):\n    pool_df = pd.read_csv(source_pool_csv)\n    pool_labels = {str(l).upper() for l in pool_df[\"label\"].unique()}\n    mn_mask = labels_upper.isin(pool_labels)\n    mn_df = df[mn_mask]\nelse:\n    pool_missing = True\n    pool_labels = set()\n    mn_df = df.iloc[0:0]\nos.makedirs(os.path.dirname(mining_gaps_parquet) or \".\", exist_ok=True)\nmn_df.to_parquet(mining_gaps_parquet, index=False)\n\n# AnomalyGen subset\nag_mask = labels_upper.isin(ANOMALYGEN_SUPPORTED)\nag_df = df[ag_mask]\nos.makedirs(os.path.dirname(anomalygen_gaps_parquet) or \".\", exist_ok=True)\nag_df.to_parquet(anomalygen_gaps_parquet, index=False)\n\n# Per-label breakdown\nsummary_lines = [\n    \"Weak-sample routing summary\",\n    f\"Total weak samples: {len(df)}\",\n    f\"Mining subset:      {len(mn_df)} -> {mining_gaps_parquet}\",\n    f\"AnomalyGen subset:  {len(ag_df)} -> {anomalygen_gaps_parquet}\",\n    \"\",\n]\nif pool_missing:\n    summary_lines.append(f\"No source pool CSV at {source_pool_csv!r}; mining subset is empty.\")\n    summary_lines.append(\"\")\nsummary_lines.append(\"Per-label breakdown (count, mining, anomalygen):\")\nlabel_counts = labels_upper.value_counts()\nfor label, count in label_counts.items():\n    in_mn = (not pool_missing) and label in pool_labels\n    in_ag = label in ANOMALYGEN_SUPPORTED\n    summary_lines.append(\n        f\"  {label}: {count} \"\n        f\"(mining={'yes' if in_mn else 'no'}, \"\n        f\"anomalygen={'yes' if in_ag else 'no'})\"\n    )\nsummary_text = \"\\n\".join(summary_lines) + \"\\n\"\n\nos.makedirs(logs_dir, exist_ok=True)\nwith open(os.path.join(logs_dir, \"routing_summary.txt\"), \"w\", encoding=\"utf-8\") as f:\n    f.write(summary_text)\nprint(summary_text.strip())\n```\n\n---\n\n## Outputs\n\nWrite everything into a timestamped folder. Any runtime packaging hook may add `routing_config\u002F` and session-capture artifacts after `Routing_Report.md` is written.\n\n```\n\u003Coutput_dir>\u002Frouting_results\u002FYYYY-MM-DD_HHMMSS\u002F\n├── Routing_Report.md           # Full routing report\n├── mining_gaps.parquet         # Subset routed to k-NN Mining\n├── anomalygen_gaps.parquet     # Subset routed to AnomalyGen (Cosmos SDG)\n├── routing_summary.txt         # Plain-text per-label breakdown\n├── routing_config\u002F             # Auto-copied by hook\n└── session log\u002Fartifacts       # Optional, runtime-dependent packaging capture\n```\n\nAt the start of the run, get the real timestamp by running `date +%Y-%m-%d_%H%M%S` in Bash. If the user specifies a custom output path, use it directly but maintain the internal layout.\n\n---\n\n## Report Structure\n\nKeep the report short (400–800 words). Routing is a deterministic decision; the value is making the decisions auditable, not narrative.\n\n```\n# VCN Routing Report: \u003CIteration \u002F Experiment Name>\n\n## 1. Verdict\n- Total weak samples in: \u003CN>\n- Mining subset:     \u003CN_mn> rows  →  `mining_gaps.parquet`\n- AnomalyGen subset: \u003CN_ag> rows  →  `anomalygen_gaps.parquet`\n- Source pool present? \u003Cyes\u002Fno — and the path>\n- One-line headline: \"\u003CX> labels routed, \u003CY> labels dropped (no module accepted)\"\n\n## 2. Inputs\n| Input | Path | Notes |\n|-------|------|-------|\n| gaps_parquet     | … | rows=\u003CN>, columns=\u003Ccol list> |\n| source_pool_csv  | … | rows=\u003CM> or \"not provided\" \u002F \"missing\" |\n\n## 3. Per-Label Routing Decisions\n| Label | Count in gaps | In source pool? | Mining? | AnomalyGen? | Routed To |\n|-------|----------------|------------------|----------|--------------|-----------|\n\n(One row per distinct label in `gaps_parquet`, uppercased. `Routed To` is one of:\n`mining only`, `anomalygen only`, `mining+anomalygen`, `neither (DROPPED)`.\nUse `neither (DROPPED)` whenever no module accepted the label. Sort by count descending.)\n\n## 4. Module-Level Summaries\n### 4.1 k-NN Mining\n- Pool labels (from source_pool_csv): \u003Clist, or \"pool missing\">\n- Labels accepted from input: \u003Clist>\n- Total rows routed: \u003CN_mn>\n- Per-label row counts: \u003Cbreakdown>\n\n### 4.2 AnomalyGen (Cosmos SDG)\n- Eligible labels (configured): PASS, EXCESS_SOLDER, MISSING, BRIDGE\n- Labels accepted from input: \u003Clist>\n- Total rows routed: \u003CN_ag>\n- Per-label row counts: \u003Cbreakdown>\n\n## 5. Dropped Labels (routed to NEITHER module)\n| Label | Count | Why dropped | Suggested fix |\n|-------|-------|-------------|----------------|\n\n(Empty table is OK and means no labels were dropped. If non-empty, every row needs a\n\"why\" — typically one of: \"not in source pool AND not in AnomalyGen supported set\",\n\"source pool missing entirely AND label not in AnomalyGen set\", \"label name doesn't\nmatch any module's expected canonicalization\".)\n\n## 6. Recommended Actions\n1. **If any labels are dropped**: seed the source pool with that label, OR extend\n   `ANOMALYGEN_SUPPORTED_LABELS` (and the AnomalyGen generator coverage).\n2. **If source pool is missing**: provide `source_pool_csv` to enable the Mining branch.\n   Without it, half of the augmentation pipeline is dark.\n3. **If AnomalyGen subset is empty**: gap analysis only surfaced labels AnomalyGen cannot\n   generate; rely on Mining for this iteration, or extend the AnomalyGen integration.\n4. **If both subsets are empty**: stop the SDA iteration. Nothing downstream can run.\n```\n\n---\n\n## Execution Order\n\n1. Run `date +%Y-%m-%d_%H%M%S` to get the timestamp; create `\u003Coutput_dir>\u002Frouting_results\u002F\u003Ctimestamp>\u002F`.\n2. Run the Python recipe (Steps 1–4) to produce `mining_gaps.parquet`, `anomalygen_gaps.parquet`, and `routing_summary.txt`. Print summary stats to stdout so the script-check hook can verify it ran.\n3. Build the per-label decision table by reading both parquets and computing the routed-to verdict per label.\n4. Write `Routing_Report.md` last — writing it triggers the packaging hook, which copies session logs and skill config alongside.\n",{"data":32,"body":44},{"name":4,"description":6,"license":23,"compatibility":33,"metadata":34,"allowed-tools":37,"tags":38},"Standalone — no external runtime requirements.",{"author":35,"version":36},"NVIDIA Corporation","0.1.0","Read Bash",[39,40,41,42,43],"data","routing","vcn","aoi","sda",{"type":45,"children":46},"root",[47,56,62,141,153,166,170,177,333,336,342,354,361,392,411,417,512,524,530,569,582,588,593,643,662,675,685,691,696,759,762,768,780,1253,1256,1262,1283,1292,1305,1308,1314,1319,1328,1331,1337,1402],{"type":48,"tag":49,"props":50,"children":52},"element","h1",{"id":51},"tao-vcn-sample-routing-skill",[53],{"type":54,"value":55},"text","TAO VCN Sample Routing Skill",{"type":48,"tag":57,"props":58,"children":59},"p",{},[60],{"type":54,"value":61},"You are the dispatcher between gap analysis and the augmentation modules in a VCN AOI SDA pipeline. Each augmentation module can only act on labels it knows how to handle:",{"type":48,"tag":63,"props":64,"children":65},"ul",{},[66,101],{"type":48,"tag":67,"props":68,"children":69},"li",{},[70,76,78,83,85,92,94,99],{"type":48,"tag":71,"props":72,"children":73},"strong",{},[74],{"type":54,"value":75},"k-NN Mining",{"type":54,"value":77}," can only mine real-image neighbors for labels that already exist in the ",{"type":48,"tag":71,"props":79,"children":80},{},[81],{"type":54,"value":82},"source pool CSV",{"type":54,"value":84},". There is no point looking for ",{"type":48,"tag":86,"props":87,"children":89},"code",{"className":88},[],[90],{"type":54,"value":91},"SHIFT",{"type":54,"value":93}," neighbors if the pool has no ",{"type":48,"tag":86,"props":95,"children":97},{"className":96},[],[98],{"type":54,"value":91},{"type":54,"value":100}," rows.",{"type":48,"tag":67,"props":102,"children":103},{},[104,109,111,117,119,125,126,132,133,139],{"type":48,"tag":71,"props":105,"children":106},{},[107],{"type":54,"value":108},"AnomalyGen",{"type":54,"value":110}," (Cosmos SDG) can only generate synthetic anomalies for the classes its inference pipeline supports: ",{"type":48,"tag":86,"props":112,"children":114},{"className":113},[],[115],{"type":54,"value":116},"PASS",{"type":54,"value":118},", ",{"type":48,"tag":86,"props":120,"children":122},{"className":121},[],[123],{"type":54,"value":124},"EXCESS_SOLDER",{"type":54,"value":118},{"type":48,"tag":86,"props":127,"children":129},{"className":128},[],[130],{"type":54,"value":131},"MISSING",{"type":54,"value":118},{"type":48,"tag":86,"props":134,"children":136},{"className":135},[],[137],{"type":54,"value":138},"BRIDGE",{"type":54,"value":140},". A weak sample with a label outside this set is unroutable to AnomalyGen.",{"type":48,"tag":57,"props":142,"children":143},{},[144,146,151],{"type":54,"value":145},"This skill runs ",{"type":48,"tag":71,"props":147,"children":148},{},[149],{"type":54,"value":150},"once per SDA iteration immediately after gap analysis",{"type":54,"value":152},". It splits the gap-analysis parquet into one filtered parquet per module so each module operates on its own eligible subset, and it writes a human-readable summary of the per-label routing decisions.",{"type":48,"tag":57,"props":154,"children":155},{},[156,158,164],{"type":54,"value":157},"The work is intentionally trivial: read a parquet, do two ",{"type":48,"tag":86,"props":159,"children":161},{"className":160},[],[162],{"type":54,"value":163},".isin(...)",{"type":54,"value":165}," filters, write two parquets, write one summary. The skill exists to make those decisions auditable — every label must show up in the summary with a yes\u002Fno verdict for each module so a downstream reviewer can spot when a label is silently dropped because no module accepted it.",{"type":48,"tag":167,"props":168,"children":169},"hr",{},[],{"type":48,"tag":171,"props":172,"children":174},"h2",{"id":173},"inputs",[175],{"type":54,"value":176},"Inputs",{"type":48,"tag":178,"props":179,"children":180},"ol",{},[181,241,262,280],{"type":48,"tag":67,"props":182,"children":183},{},[184,193,195,201,203,209,211,217,218,224,226,232,233,239],{"type":48,"tag":71,"props":185,"children":186},{},[187],{"type":48,"tag":86,"props":188,"children":190},{"className":189},[],[191],{"type":54,"value":192},"gaps_parquet",{"type":54,"value":194}," — the gap-analysis output (typically ",{"type":48,"tag":86,"props":196,"children":198},{"className":197},[],[199],{"type":54,"value":200},"\u003Cexp_dir>\u002Frca_results\u002F\u003Ctimestamp>\u002Fgaps.parquet",{"type":54,"value":202}," from ",{"type":48,"tag":86,"props":204,"children":206},{"className":205},[],[207],{"type":54,"value":208},"tao-analyze-gaps-visual-changenet",{"type":54,"value":210},"). Required columns: ",{"type":48,"tag":86,"props":212,"children":214},{"className":213},[],[215],{"type":54,"value":216},"filepath",{"type":54,"value":118},{"type":48,"tag":86,"props":219,"children":221},{"className":220},[],[222],{"type":54,"value":223},"label",{"type":54,"value":225},". Other columns (",{"type":48,"tag":86,"props":227,"children":229},{"className":228},[],[230],{"type":54,"value":231},"siamese_score",{"type":54,"value":118},{"type":48,"tag":86,"props":234,"children":236},{"className":235},[],[237],{"type":54,"value":238},"weakness",{"type":54,"value":240},") are preserved verbatim.",{"type":48,"tag":67,"props":242,"children":243},{},[244,253,255,260],{"type":48,"tag":71,"props":245,"children":246},{},[247],{"type":48,"tag":86,"props":248,"children":250},{"className":249},[],[251],{"type":54,"value":252},"source_pool_csv",{"type":54,"value":254}," — VCN-format mining source pool CSV with a ",{"type":48,"tag":86,"props":256,"children":258},{"className":257},[],[259],{"type":54,"value":223},{"type":54,"value":261}," column. Empty string or non-existent path is allowed; the mining subset will simply be empty in that case.",{"type":48,"tag":67,"props":263,"children":264},{},[265,270,272,278],{"type":48,"tag":71,"props":266,"children":267},{},[268],{"type":54,"value":269},"Output directory",{"type":54,"value":271}," — where the two routed parquets, the summary, and the report are written. Default: a timestamped folder under the gap-analysis result directory: ",{"type":48,"tag":86,"props":273,"children":275},{"className":274},[],[276],{"type":54,"value":277},"\u003Crca_result_dir>\u002Frouting_results\u002F\u003Ctimestamp>\u002F",{"type":54,"value":279},".",{"type":48,"tag":67,"props":281,"children":282},{},[283,292,294,300,302,308,310,315,317,323,325,331],{"type":48,"tag":71,"props":284,"children":285},{},[286],{"type":48,"tag":86,"props":287,"children":289},{"className":288},[],[290],{"type":54,"value":291},"anomalygen_supported_labels",{"type":54,"value":293}," ",{"type":48,"tag":295,"props":296,"children":297},"em",{},[298],{"type":54,"value":299},"(optional)",{"type":54,"value":301}," — override the default AnomalyGen-eligible label set. Default: ",{"type":48,"tag":86,"props":303,"children":305},{"className":304},[],[306],{"type":54,"value":307},"{\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}",{"type":54,"value":309},". ",{"type":48,"tag":71,"props":311,"children":312},{},[313],{"type":54,"value":314},"Warning:",{"type":54,"value":316}," This must stay in sync with ",{"type":48,"tag":86,"props":318,"children":320},{"className":319},[],[321],{"type":54,"value":322},"ANOMALYGEN_SUPPORTED_LABELS",{"type":54,"value":324}," in ",{"type":48,"tag":86,"props":326,"children":328},{"className":327},[],[329],{"type":54,"value":330},"mdo-kratos-workflows\u002Fpipelines\u002Fsda\u002Frouting.py",{"type":54,"value":332}," and the AnomalyGen integration's actual generator coverage. Adding a new defect class to AnomalyGen means adding it here too.",{"type":48,"tag":167,"props":334,"children":335},{},[],{"type":48,"tag":171,"props":337,"children":339},{"id":338},"method",[340],{"type":54,"value":341},"Method",{"type":48,"tag":57,"props":343,"children":344},{},[345,347,352],{"type":54,"value":346},"The whole skill is two ",{"type":48,"tag":86,"props":348,"children":350},{"className":349},[],[351],{"type":54,"value":163},{"type":54,"value":353}," masks against the uppercased label column.",{"type":48,"tag":355,"props":356,"children":358},"h3",{"id":357},"step-1-load-and-uppercase",[359],{"type":54,"value":360},"Step 1 — Load and uppercase",{"type":48,"tag":362,"props":363,"children":368},"pre",{"className":364,"code":365,"language":366,"meta":367,"style":367},"language-python shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","df = pd.read_parquet(gaps_parquet)\nlabels_upper = df[\"label\"].astype(str).str.upper()\n","python","",[369],{"type":48,"tag":86,"props":370,"children":371},{"__ignoreMap":367},[372,383],{"type":48,"tag":373,"props":374,"children":377},"span",{"class":375,"line":376},"line",1,[378],{"type":48,"tag":373,"props":379,"children":380},{},[381],{"type":54,"value":382},"df = pd.read_parquet(gaps_parquet)\n",{"type":48,"tag":373,"props":384,"children":386},{"class":375,"line":385},2,[387],{"type":48,"tag":373,"props":388,"children":389},{},[390],{"type":54,"value":391},"labels_upper = df[\"label\"].astype(str).str.upper()\n",{"type":48,"tag":57,"props":393,"children":394},{},[395,397,402,404,409],{"type":54,"value":396},"The match is ",{"type":48,"tag":71,"props":398,"children":399},{},[400],{"type":54,"value":401},"case-insensitive",{"type":54,"value":403}," for both module checks. The original ",{"type":48,"tag":86,"props":405,"children":407},{"className":406},[],[408],{"type":54,"value":223},{"type":54,"value":410}," column is preserved unchanged in the output parquets — only the comparison key is uppercased.",{"type":48,"tag":355,"props":412,"children":414},{"id":413},"step-2-mining-subset",[415],{"type":54,"value":416},"Step 2 — Mining subset",{"type":48,"tag":362,"props":418,"children":420},{"className":364,"code":419,"language":366,"meta":367,"style":367},"if source_pool_csv and os.path.isfile(source_pool_csv):\n    pool_df = pd.read_csv(source_pool_csv)\n    pool_labels = {str(l).upper() for l in pool_df[\"label\"].unique()}\n    mn_mask = labels_upper.isin(pool_labels)\n    mn_df = df[mn_mask]\nelse:\n    pool_missing = True\n    pool_labels = set()\n    mn_df = df.iloc[0:0]   # empty, but with the same schema\nmn_df.to_parquet(mining_gaps_parquet, index=False)\n",[421],{"type":48,"tag":86,"props":422,"children":423},{"__ignoreMap":367},[424,432,440,449,458,467,476,485,494,503],{"type":48,"tag":373,"props":425,"children":426},{"class":375,"line":376},[427],{"type":48,"tag":373,"props":428,"children":429},{},[430],{"type":54,"value":431},"if source_pool_csv and os.path.isfile(source_pool_csv):\n",{"type":48,"tag":373,"props":433,"children":434},{"class":375,"line":385},[435],{"type":48,"tag":373,"props":436,"children":437},{},[438],{"type":54,"value":439},"    pool_df = pd.read_csv(source_pool_csv)\n",{"type":48,"tag":373,"props":441,"children":443},{"class":375,"line":442},3,[444],{"type":48,"tag":373,"props":445,"children":446},{},[447],{"type":54,"value":448},"    pool_labels = {str(l).upper() for l in pool_df[\"label\"].unique()}\n",{"type":48,"tag":373,"props":450,"children":452},{"class":375,"line":451},4,[453],{"type":48,"tag":373,"props":454,"children":455},{},[456],{"type":54,"value":457},"    mn_mask = labels_upper.isin(pool_labels)\n",{"type":48,"tag":373,"props":459,"children":461},{"class":375,"line":460},5,[462],{"type":48,"tag":373,"props":463,"children":464},{},[465],{"type":54,"value":466},"    mn_df = df[mn_mask]\n",{"type":48,"tag":373,"props":468,"children":470},{"class":375,"line":469},6,[471],{"type":48,"tag":373,"props":472,"children":473},{},[474],{"type":54,"value":475},"else:\n",{"type":48,"tag":373,"props":477,"children":479},{"class":375,"line":478},7,[480],{"type":48,"tag":373,"props":481,"children":482},{},[483],{"type":54,"value":484},"    pool_missing = True\n",{"type":48,"tag":373,"props":486,"children":488},{"class":375,"line":487},8,[489],{"type":48,"tag":373,"props":490,"children":491},{},[492],{"type":54,"value":493},"    pool_labels = set()\n",{"type":48,"tag":373,"props":495,"children":497},{"class":375,"line":496},9,[498],{"type":48,"tag":373,"props":499,"children":500},{},[501],{"type":54,"value":502},"    mn_df = df.iloc[0:0]   # empty, but with the same schema\n",{"type":48,"tag":373,"props":504,"children":506},{"class":375,"line":505},10,[507],{"type":48,"tag":373,"props":508,"children":509},{},[510],{"type":54,"value":511},"mn_df.to_parquet(mining_gaps_parquet, index=False)\n",{"type":48,"tag":57,"props":513,"children":514},{},[515,517,522],{"type":54,"value":516},"If the pool CSV is missing or empty, the mining subset is an empty DataFrame ",{"type":48,"tag":71,"props":518,"children":519},{},[520],{"type":54,"value":521},"with the same columns as the input",{"type":54,"value":523}," so downstream readers don't crash on schema mismatch. Flag this case in the summary.",{"type":48,"tag":355,"props":525,"children":527},{"id":526},"step-3-anomalygen-subset",[528],{"type":54,"value":529},"Step 3 — AnomalyGen subset",{"type":48,"tag":362,"props":531,"children":533},{"className":364,"code":532,"language":366,"meta":367,"style":367},"ANOMALYGEN_SUPPORTED = {\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}\nag_mask = labels_upper.isin(ANOMALYGEN_SUPPORTED)\nag_df = df[ag_mask]\nag_df.to_parquet(anomalygen_gaps_parquet, index=False)\n",[534],{"type":48,"tag":86,"props":535,"children":536},{"__ignoreMap":367},[537,545,553,561],{"type":48,"tag":373,"props":538,"children":539},{"class":375,"line":376},[540],{"type":48,"tag":373,"props":541,"children":542},{},[543],{"type":54,"value":544},"ANOMALYGEN_SUPPORTED = {\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}\n",{"type":48,"tag":373,"props":546,"children":547},{"class":375,"line":385},[548],{"type":48,"tag":373,"props":549,"children":550},{},[551],{"type":54,"value":552},"ag_mask = labels_upper.isin(ANOMALYGEN_SUPPORTED)\n",{"type":48,"tag":373,"props":554,"children":555},{"class":375,"line":442},[556],{"type":48,"tag":373,"props":557,"children":558},{},[559],{"type":54,"value":560},"ag_df = df[ag_mask]\n",{"type":48,"tag":373,"props":562,"children":563},{"class":375,"line":451},[564],{"type":48,"tag":373,"props":565,"children":566},{},[567],{"type":54,"value":568},"ag_df.to_parquet(anomalygen_gaps_parquet, index=False)\n",{"type":48,"tag":57,"props":570,"children":571},{},[572,574,580],{"type":54,"value":573},"Rows whose label is in the AnomalyGen-supported set are written verbatim to ",{"type":48,"tag":86,"props":575,"children":577},{"className":576},[],[578],{"type":54,"value":579},"anomalygen_gaps.parquet",{"type":54,"value":581},". The schema matches the input parquet exactly — downstream AnomalyGen (Cosmos SDG) needs no other changes.",{"type":48,"tag":355,"props":583,"children":585},{"id":584},"step-4-per-label-routing-breakdown",[586],{"type":54,"value":587},"Step 4 — Per-label routing breakdown",{"type":48,"tag":57,"props":589,"children":590},{},[591],{"type":54,"value":592},"For every distinct label in the input gaps parquet (uppercased), record:",{"type":48,"tag":63,"props":594,"children":595},{},[596,607,626],{"type":48,"tag":67,"props":597,"children":598},{},[599,605],{"type":48,"tag":86,"props":600,"children":602},{"className":601},[],[603],{"type":54,"value":604},"count",{"type":54,"value":606}," — how many rows have this label",{"type":48,"tag":67,"props":608,"children":609},{},[610,616,618,624],{"type":48,"tag":86,"props":611,"children":613},{"className":612},[],[614],{"type":54,"value":615},"mining",{"type":54,"value":617}," — yes if the label is in ",{"type":48,"tag":86,"props":619,"children":621},{"className":620},[],[622],{"type":54,"value":623},"pool_labels",{"type":54,"value":625},", otherwise no",{"type":48,"tag":67,"props":627,"children":628},{},[629,635,636,642],{"type":48,"tag":86,"props":630,"children":632},{"className":631},[],[633],{"type":54,"value":634},"anomalygen",{"type":54,"value":617},{"type":48,"tag":86,"props":637,"children":639},{"className":638},[],[640],{"type":54,"value":641},"ANOMALYGEN_SUPPORTED",{"type":54,"value":625},{"type":48,"tag":57,"props":644,"children":645},{},[646,648,653,655,660],{"type":54,"value":647},"A label can route to ",{"type":48,"tag":71,"props":649,"children":650},{},[651],{"type":54,"value":652},"both",{"type":54,"value":654}," modules (e.g. PASS rows route to AnomalyGen, and if the source pool also contains PASS rows they route to Mining too). A label can also route to ",{"type":48,"tag":71,"props":656,"children":657},{},[658],{"type":54,"value":659},"none",{"type":54,"value":661}," — flag those, since they are silently dropped and may signal a configuration mismatch.",{"type":48,"tag":57,"props":663,"children":664},{},[665,667,673],{"type":54,"value":666},"Write the breakdown to ",{"type":48,"tag":86,"props":668,"children":670},{"className":669},[],[671],{"type":54,"value":672},"routing_summary.txt",{"type":54,"value":674},". The format mirrors the reference component exactly:",{"type":48,"tag":362,"props":676,"children":680},{"className":677,"code":679,"language":54},[678],"language-text","Weak-sample routing summary\nTotal weak samples: \u003CN>\nMining subset:      \u003CN_mn> -> \u003Cmining_gaps_parquet>\nAnomalyGen subset:  \u003CN_ag> -> \u003Canomalygen_gaps_parquet>\n\n[If pool missing:]\nNo source pool CSV at '\u003Cpath>'; mining subset is empty.\n\nPer-label breakdown (count, mining, anomalygen):\n  PASS: 50 (mining=yes, anomalygen=yes)\n  MISSING: 32 (mining=no, anomalygen=yes)\n  SHIFT: 14 (mining=yes, anomalygen=no)\n  EXCESS_SOLDER: 9 (mining=yes, anomalygen=yes)\n  ...\n",[681],{"type":48,"tag":86,"props":682,"children":683},{"__ignoreMap":367},[684],{"type":54,"value":679},{"type":48,"tag":355,"props":686,"children":688},{"id":687},"step-5-sanity-checks",[689],{"type":54,"value":690},"Step 5 — Sanity checks",{"type":48,"tag":57,"props":692,"children":693},{},[694],{"type":54,"value":695},"After both subsets are written, verify:",{"type":48,"tag":63,"props":697,"children":698},{},[699,725,746],{"type":48,"tag":67,"props":700,"children":701},{},[702,704,709,711,717,719,724],{"type":54,"value":703},"The sum of subset sizes is ",{"type":48,"tag":295,"props":705,"children":706},{},[707],{"type":54,"value":708},"not",{"type":54,"value":710}," required to equal ",{"type":48,"tag":86,"props":712,"children":714},{"className":713},[],[715],{"type":54,"value":716},"len(df)",{"type":54,"value":718}," — overlap is allowed (a label can route to both modules). What matters is that ",{"type":48,"tag":71,"props":720,"children":721},{},[722],{"type":54,"value":723},"every input row appears in at least one subset, OR appears in the \"none\" list with an explicit reason",{"type":54,"value":279},{"type":48,"tag":67,"props":726,"children":727},{},[728,730,736,738,744],{"type":54,"value":729},"If ",{"type":48,"tag":86,"props":731,"children":733},{"className":732},[],[734],{"type":54,"value":735},"len(mn_df) == 0",{"type":54,"value":737}," and ",{"type":48,"tag":86,"props":739,"children":741},{"className":740},[],[742],{"type":54,"value":743},"len(ag_df) == 0",{"type":54,"value":745},", something is wrong — flag prominently in the report.",{"type":48,"tag":67,"props":747,"children":748},{},[749,751,757],{"type":54,"value":750},"If an entire label group routes to no module, the ",{"type":48,"tag":86,"props":752,"children":754},{"className":753},[],[755],{"type":54,"value":756},"Recommended Actions",{"type":54,"value":758}," section must call this out so the user can either seed the source pool with that label or extend AnomalyGen's supported set.",{"type":48,"tag":167,"props":760,"children":761},{},[],{"type":48,"tag":171,"props":763,"children":765},{"id":764},"reference-python-recipe",[766],{"type":54,"value":767},"Reference Python Recipe",{"type":48,"tag":57,"props":769,"children":770},{},[771,773,778],{"type":54,"value":772},"This is the exact computation, lifted from ",{"type":48,"tag":86,"props":774,"children":776},{"className":775},[],[777],{"type":54,"value":330},{"type":54,"value":779},". Run as a single Python script via Bash; it produces every artifact except the report.",{"type":48,"tag":362,"props":781,"children":783},{"className":364,"code":782,"language":366,"meta":367,"style":367},"import os\nimport pandas as pd\n\nANOMALYGEN_SUPPORTED = {\"PASS\", \"EXCESS_SOLDER\", \"MISSING\", \"BRIDGE\"}\n\ndf = pd.read_parquet(gaps_parquet)\nlabels_upper = df[\"label\"].astype(str).str.upper()\n\n# Mining subset\npool_missing = False\nif source_pool_csv and os.path.isfile(source_pool_csv):\n    pool_df = pd.read_csv(source_pool_csv)\n    pool_labels = {str(l).upper() for l in pool_df[\"label\"].unique()}\n    mn_mask = labels_upper.isin(pool_labels)\n    mn_df = df[mn_mask]\nelse:\n    pool_missing = True\n    pool_labels = set()\n    mn_df = df.iloc[0:0]\nos.makedirs(os.path.dirname(mining_gaps_parquet) or \".\", exist_ok=True)\nmn_df.to_parquet(mining_gaps_parquet, index=False)\n\n# AnomalyGen subset\nag_mask = labels_upper.isin(ANOMALYGEN_SUPPORTED)\nag_df = df[ag_mask]\nos.makedirs(os.path.dirname(anomalygen_gaps_parquet) or \".\", exist_ok=True)\nag_df.to_parquet(anomalygen_gaps_parquet, index=False)\n\n# Per-label breakdown\nsummary_lines = [\n    \"Weak-sample routing summary\",\n    f\"Total weak samples: {len(df)}\",\n    f\"Mining subset:      {len(mn_df)} -> {mining_gaps_parquet}\",\n    f\"AnomalyGen subset:  {len(ag_df)} -> {anomalygen_gaps_parquet}\",\n    \"\",\n]\nif pool_missing:\n    summary_lines.append(f\"No source pool CSV at {source_pool_csv!r}; mining subset is empty.\")\n    summary_lines.append(\"\")\nsummary_lines.append(\"Per-label breakdown (count, mining, anomalygen):\")\nlabel_counts = labels_upper.value_counts()\nfor label, count in label_counts.items():\n    in_mn = (not pool_missing) and label in pool_labels\n    in_ag = label in ANOMALYGEN_SUPPORTED\n    summary_lines.append(\n        f\"  {label}: {count} \"\n        f\"(mining={'yes' if in_mn else 'no'}, \"\n        f\"anomalygen={'yes' if in_ag else 'no'})\"\n    )\nsummary_text = \"\\n\".join(summary_lines) + \"\\n\"\n\nos.makedirs(logs_dir, exist_ok=True)\nwith open(os.path.join(logs_dir, \"routing_summary.txt\"), \"w\", encoding=\"utf-8\") as f:\n    f.write(summary_text)\nprint(summary_text.strip())\n",[784],{"type":48,"tag":86,"props":785,"children":786},{"__ignoreMap":367},[787,795,803,812,819,826,833,840,847,855,863,871,879,887,895,903,911,919,927,936,945,953,961,970,978,986,995,1003,1011,1020,1029,1038,1047,1056,1065,1074,1083,1092,1101,1110,1119,1128,1137,1146,1155,1164,1173,1182,1191,1200,1209,1217,1226,1235,1244],{"type":48,"tag":373,"props":788,"children":789},{"class":375,"line":376},[790],{"type":48,"tag":373,"props":791,"children":792},{},[793],{"type":54,"value":794},"import os\n",{"type":48,"tag":373,"props":796,"children":797},{"class":375,"line":385},[798],{"type":48,"tag":373,"props":799,"children":800},{},[801],{"type":54,"value":802},"import pandas as pd\n",{"type":48,"tag":373,"props":804,"children":805},{"class":375,"line":442},[806],{"type":48,"tag":373,"props":807,"children":809},{"emptyLinePlaceholder":808},true,[810],{"type":54,"value":811},"\n",{"type":48,"tag":373,"props":813,"children":814},{"class":375,"line":451},[815],{"type":48,"tag":373,"props":816,"children":817},{},[818],{"type":54,"value":544},{"type":48,"tag":373,"props":820,"children":821},{"class":375,"line":460},[822],{"type":48,"tag":373,"props":823,"children":824},{"emptyLinePlaceholder":808},[825],{"type":54,"value":811},{"type":48,"tag":373,"props":827,"children":828},{"class":375,"line":469},[829],{"type":48,"tag":373,"props":830,"children":831},{},[832],{"type":54,"value":382},{"type":48,"tag":373,"props":834,"children":835},{"class":375,"line":478},[836],{"type":48,"tag":373,"props":837,"children":838},{},[839],{"type":54,"value":391},{"type":48,"tag":373,"props":841,"children":842},{"class":375,"line":487},[843],{"type":48,"tag":373,"props":844,"children":845},{"emptyLinePlaceholder":808},[846],{"type":54,"value":811},{"type":48,"tag":373,"props":848,"children":849},{"class":375,"line":496},[850],{"type":48,"tag":373,"props":851,"children":852},{},[853],{"type":54,"value":854},"# Mining subset\n",{"type":48,"tag":373,"props":856,"children":857},{"class":375,"line":505},[858],{"type":48,"tag":373,"props":859,"children":860},{},[861],{"type":54,"value":862},"pool_missing = False\n",{"type":48,"tag":373,"props":864,"children":866},{"class":375,"line":865},11,[867],{"type":48,"tag":373,"props":868,"children":869},{},[870],{"type":54,"value":431},{"type":48,"tag":373,"props":872,"children":874},{"class":375,"line":873},12,[875],{"type":48,"tag":373,"props":876,"children":877},{},[878],{"type":54,"value":439},{"type":48,"tag":373,"props":880,"children":882},{"class":375,"line":881},13,[883],{"type":48,"tag":373,"props":884,"children":885},{},[886],{"type":54,"value":448},{"type":48,"tag":373,"props":888,"children":890},{"class":375,"line":889},14,[891],{"type":48,"tag":373,"props":892,"children":893},{},[894],{"type":54,"value":457},{"type":48,"tag":373,"props":896,"children":898},{"class":375,"line":897},15,[899],{"type":48,"tag":373,"props":900,"children":901},{},[902],{"type":54,"value":466},{"type":48,"tag":373,"props":904,"children":906},{"class":375,"line":905},16,[907],{"type":48,"tag":373,"props":908,"children":909},{},[910],{"type":54,"value":475},{"type":48,"tag":373,"props":912,"children":914},{"class":375,"line":913},17,[915],{"type":48,"tag":373,"props":916,"children":917},{},[918],{"type":54,"value":484},{"type":48,"tag":373,"props":920,"children":922},{"class":375,"line":921},18,[923],{"type":48,"tag":373,"props":924,"children":925},{},[926],{"type":54,"value":493},{"type":48,"tag":373,"props":928,"children":930},{"class":375,"line":929},19,[931],{"type":48,"tag":373,"props":932,"children":933},{},[934],{"type":54,"value":935},"    mn_df = df.iloc[0:0]\n",{"type":48,"tag":373,"props":937,"children":939},{"class":375,"line":938},20,[940],{"type":48,"tag":373,"props":941,"children":942},{},[943],{"type":54,"value":944},"os.makedirs(os.path.dirname(mining_gaps_parquet) or \".\", exist_ok=True)\n",{"type":48,"tag":373,"props":946,"children":948},{"class":375,"line":947},21,[949],{"type":48,"tag":373,"props":950,"children":951},{},[952],{"type":54,"value":511},{"type":48,"tag":373,"props":954,"children":956},{"class":375,"line":955},22,[957],{"type":48,"tag":373,"props":958,"children":959},{"emptyLinePlaceholder":808},[960],{"type":54,"value":811},{"type":48,"tag":373,"props":962,"children":964},{"class":375,"line":963},23,[965],{"type":48,"tag":373,"props":966,"children":967},{},[968],{"type":54,"value":969},"# AnomalyGen subset\n",{"type":48,"tag":373,"props":971,"children":973},{"class":375,"line":972},24,[974],{"type":48,"tag":373,"props":975,"children":976},{},[977],{"type":54,"value":552},{"type":48,"tag":373,"props":979,"children":981},{"class":375,"line":980},25,[982],{"type":48,"tag":373,"props":983,"children":984},{},[985],{"type":54,"value":560},{"type":48,"tag":373,"props":987,"children":989},{"class":375,"line":988},26,[990],{"type":48,"tag":373,"props":991,"children":992},{},[993],{"type":54,"value":994},"os.makedirs(os.path.dirname(anomalygen_gaps_parquet) or \".\", exist_ok=True)\n",{"type":48,"tag":373,"props":996,"children":998},{"class":375,"line":997},27,[999],{"type":48,"tag":373,"props":1000,"children":1001},{},[1002],{"type":54,"value":568},{"type":48,"tag":373,"props":1004,"children":1006},{"class":375,"line":1005},28,[1007],{"type":48,"tag":373,"props":1008,"children":1009},{"emptyLinePlaceholder":808},[1010],{"type":54,"value":811},{"type":48,"tag":373,"props":1012,"children":1014},{"class":375,"line":1013},29,[1015],{"type":48,"tag":373,"props":1016,"children":1017},{},[1018],{"type":54,"value":1019},"# Per-label breakdown\n",{"type":48,"tag":373,"props":1021,"children":1023},{"class":375,"line":1022},30,[1024],{"type":48,"tag":373,"props":1025,"children":1026},{},[1027],{"type":54,"value":1028},"summary_lines = [\n",{"type":48,"tag":373,"props":1030,"children":1032},{"class":375,"line":1031},31,[1033],{"type":48,"tag":373,"props":1034,"children":1035},{},[1036],{"type":54,"value":1037},"    \"Weak-sample routing summary\",\n",{"type":48,"tag":373,"props":1039,"children":1041},{"class":375,"line":1040},32,[1042],{"type":48,"tag":373,"props":1043,"children":1044},{},[1045],{"type":54,"value":1046},"    f\"Total weak samples: {len(df)}\",\n",{"type":48,"tag":373,"props":1048,"children":1050},{"class":375,"line":1049},33,[1051],{"type":48,"tag":373,"props":1052,"children":1053},{},[1054],{"type":54,"value":1055},"    f\"Mining subset:      {len(mn_df)} -> {mining_gaps_parquet}\",\n",{"type":48,"tag":373,"props":1057,"children":1059},{"class":375,"line":1058},34,[1060],{"type":48,"tag":373,"props":1061,"children":1062},{},[1063],{"type":54,"value":1064},"    f\"AnomalyGen subset:  {len(ag_df)} -> {anomalygen_gaps_parquet}\",\n",{"type":48,"tag":373,"props":1066,"children":1068},{"class":375,"line":1067},35,[1069],{"type":48,"tag":373,"props":1070,"children":1071},{},[1072],{"type":54,"value":1073},"    \"\",\n",{"type":48,"tag":373,"props":1075,"children":1077},{"class":375,"line":1076},36,[1078],{"type":48,"tag":373,"props":1079,"children":1080},{},[1081],{"type":54,"value":1082},"]\n",{"type":48,"tag":373,"props":1084,"children":1086},{"class":375,"line":1085},37,[1087],{"type":48,"tag":373,"props":1088,"children":1089},{},[1090],{"type":54,"value":1091},"if pool_missing:\n",{"type":48,"tag":373,"props":1093,"children":1095},{"class":375,"line":1094},38,[1096],{"type":48,"tag":373,"props":1097,"children":1098},{},[1099],{"type":54,"value":1100},"    summary_lines.append(f\"No source pool CSV at {source_pool_csv!r}; mining subset is empty.\")\n",{"type":48,"tag":373,"props":1102,"children":1104},{"class":375,"line":1103},39,[1105],{"type":48,"tag":373,"props":1106,"children":1107},{},[1108],{"type":54,"value":1109},"    summary_lines.append(\"\")\n",{"type":48,"tag":373,"props":1111,"children":1113},{"class":375,"line":1112},40,[1114],{"type":48,"tag":373,"props":1115,"children":1116},{},[1117],{"type":54,"value":1118},"summary_lines.append(\"Per-label breakdown (count, mining, anomalygen):\")\n",{"type":48,"tag":373,"props":1120,"children":1122},{"class":375,"line":1121},41,[1123],{"type":48,"tag":373,"props":1124,"children":1125},{},[1126],{"type":54,"value":1127},"label_counts = labels_upper.value_counts()\n",{"type":48,"tag":373,"props":1129,"children":1131},{"class":375,"line":1130},42,[1132],{"type":48,"tag":373,"props":1133,"children":1134},{},[1135],{"type":54,"value":1136},"for label, count in label_counts.items():\n",{"type":48,"tag":373,"props":1138,"children":1140},{"class":375,"line":1139},43,[1141],{"type":48,"tag":373,"props":1142,"children":1143},{},[1144],{"type":54,"value":1145},"    in_mn = (not pool_missing) and label in pool_labels\n",{"type":48,"tag":373,"props":1147,"children":1149},{"class":375,"line":1148},44,[1150],{"type":48,"tag":373,"props":1151,"children":1152},{},[1153],{"type":54,"value":1154},"    in_ag = label in ANOMALYGEN_SUPPORTED\n",{"type":48,"tag":373,"props":1156,"children":1158},{"class":375,"line":1157},45,[1159],{"type":48,"tag":373,"props":1160,"children":1161},{},[1162],{"type":54,"value":1163},"    summary_lines.append(\n",{"type":48,"tag":373,"props":1165,"children":1167},{"class":375,"line":1166},46,[1168],{"type":48,"tag":373,"props":1169,"children":1170},{},[1171],{"type":54,"value":1172},"        f\"  {label}: {count} \"\n",{"type":48,"tag":373,"props":1174,"children":1176},{"class":375,"line":1175},47,[1177],{"type":48,"tag":373,"props":1178,"children":1179},{},[1180],{"type":54,"value":1181},"        f\"(mining={'yes' if in_mn else 'no'}, \"\n",{"type":48,"tag":373,"props":1183,"children":1185},{"class":375,"line":1184},48,[1186],{"type":48,"tag":373,"props":1187,"children":1188},{},[1189],{"type":54,"value":1190},"        f\"anomalygen={'yes' if in_ag else 'no'})\"\n",{"type":48,"tag":373,"props":1192,"children":1194},{"class":375,"line":1193},49,[1195],{"type":48,"tag":373,"props":1196,"children":1197},{},[1198],{"type":54,"value":1199},"    )\n",{"type":48,"tag":373,"props":1201,"children":1203},{"class":375,"line":1202},50,[1204],{"type":48,"tag":373,"props":1205,"children":1206},{},[1207],{"type":54,"value":1208},"summary_text = \"\\n\".join(summary_lines) + \"\\n\"\n",{"type":48,"tag":373,"props":1210,"children":1212},{"class":375,"line":1211},51,[1213],{"type":48,"tag":373,"props":1214,"children":1215},{"emptyLinePlaceholder":808},[1216],{"type":54,"value":811},{"type":48,"tag":373,"props":1218,"children":1220},{"class":375,"line":1219},52,[1221],{"type":48,"tag":373,"props":1222,"children":1223},{},[1224],{"type":54,"value":1225},"os.makedirs(logs_dir, exist_ok=True)\n",{"type":48,"tag":373,"props":1227,"children":1229},{"class":375,"line":1228},53,[1230],{"type":48,"tag":373,"props":1231,"children":1232},{},[1233],{"type":54,"value":1234},"with open(os.path.join(logs_dir, \"routing_summary.txt\"), \"w\", encoding=\"utf-8\") as f:\n",{"type":48,"tag":373,"props":1236,"children":1238},{"class":375,"line":1237},54,[1239],{"type":48,"tag":373,"props":1240,"children":1241},{},[1242],{"type":54,"value":1243},"    f.write(summary_text)\n",{"type":48,"tag":373,"props":1245,"children":1247},{"class":375,"line":1246},55,[1248],{"type":48,"tag":373,"props":1249,"children":1250},{},[1251],{"type":54,"value":1252},"print(summary_text.strip())\n",{"type":48,"tag":167,"props":1254,"children":1255},{},[],{"type":48,"tag":171,"props":1257,"children":1259},{"id":1258},"outputs",[1260],{"type":54,"value":1261},"Outputs",{"type":48,"tag":57,"props":1263,"children":1264},{},[1265,1267,1273,1275,1281],{"type":54,"value":1266},"Write everything into a timestamped folder. Any runtime packaging hook may add ",{"type":48,"tag":86,"props":1268,"children":1270},{"className":1269},[],[1271],{"type":54,"value":1272},"routing_config\u002F",{"type":54,"value":1274}," and session-capture artifacts after ",{"type":48,"tag":86,"props":1276,"children":1278},{"className":1277},[],[1279],{"type":54,"value":1280},"Routing_Report.md",{"type":54,"value":1282}," is written.",{"type":48,"tag":362,"props":1284,"children":1287},{"className":1285,"code":1286,"language":54},[678],"\u003Coutput_dir>\u002Frouting_results\u002FYYYY-MM-DD_HHMMSS\u002F\n├── Routing_Report.md           # Full routing report\n├── mining_gaps.parquet         # Subset routed to k-NN Mining\n├── anomalygen_gaps.parquet     # Subset routed to AnomalyGen (Cosmos SDG)\n├── routing_summary.txt         # Plain-text per-label breakdown\n├── routing_config\u002F             # Auto-copied by hook\n└── session log\u002Fartifacts       # Optional, runtime-dependent packaging capture\n",[1288],{"type":48,"tag":86,"props":1289,"children":1290},{"__ignoreMap":367},[1291],{"type":54,"value":1286},{"type":48,"tag":57,"props":1293,"children":1294},{},[1295,1297,1303],{"type":54,"value":1296},"At the start of the run, get the real timestamp by running ",{"type":48,"tag":86,"props":1298,"children":1300},{"className":1299},[],[1301],{"type":54,"value":1302},"date +%Y-%m-%d_%H%M%S",{"type":54,"value":1304}," in Bash. If the user specifies a custom output path, use it directly but maintain the internal layout.",{"type":48,"tag":167,"props":1306,"children":1307},{},[],{"type":48,"tag":171,"props":1309,"children":1311},{"id":1310},"report-structure",[1312],{"type":54,"value":1313},"Report Structure",{"type":48,"tag":57,"props":1315,"children":1316},{},[1317],{"type":54,"value":1318},"Keep the report short (400–800 words). Routing is a deterministic decision; the value is making the decisions auditable, not narrative.",{"type":48,"tag":362,"props":1320,"children":1323},{"className":1321,"code":1322,"language":54},[678],"# VCN Routing Report: \u003CIteration \u002F Experiment Name>\n\n## 1. Verdict\n- Total weak samples in: \u003CN>\n- Mining subset:     \u003CN_mn> rows  →  `mining_gaps.parquet`\n- AnomalyGen subset: \u003CN_ag> rows  →  `anomalygen_gaps.parquet`\n- Source pool present? \u003Cyes\u002Fno — and the path>\n- One-line headline: \"\u003CX> labels routed, \u003CY> labels dropped (no module accepted)\"\n\n## 2. Inputs\n| Input | Path | Notes |\n|-------|------|-------|\n| gaps_parquet     | … | rows=\u003CN>, columns=\u003Ccol list> |\n| source_pool_csv  | … | rows=\u003CM> or \"not provided\" \u002F \"missing\" |\n\n## 3. Per-Label Routing Decisions\n| Label | Count in gaps | In source pool? | Mining? | AnomalyGen? | Routed To |\n|-------|----------------|------------------|----------|--------------|-----------|\n\n(One row per distinct label in `gaps_parquet`, uppercased. `Routed To` is one of:\n`mining only`, `anomalygen only`, `mining+anomalygen`, `neither (DROPPED)`.\nUse `neither (DROPPED)` whenever no module accepted the label. Sort by count descending.)\n\n## 4. Module-Level Summaries\n### 4.1 k-NN Mining\n- Pool labels (from source_pool_csv): \u003Clist, or \"pool missing\">\n- Labels accepted from input: \u003Clist>\n- Total rows routed: \u003CN_mn>\n- Per-label row counts: \u003Cbreakdown>\n\n### 4.2 AnomalyGen (Cosmos SDG)\n- Eligible labels (configured): PASS, EXCESS_SOLDER, MISSING, BRIDGE\n- Labels accepted from input: \u003Clist>\n- Total rows routed: \u003CN_ag>\n- Per-label row counts: \u003Cbreakdown>\n\n## 5. Dropped Labels (routed to NEITHER module)\n| Label | Count | Why dropped | Suggested fix |\n|-------|-------|-------------|----------------|\n\n(Empty table is OK and means no labels were dropped. If non-empty, every row needs a\n\"why\" — typically one of: \"not in source pool AND not in AnomalyGen supported set\",\n\"source pool missing entirely AND label not in AnomalyGen set\", \"label name doesn't\nmatch any module's expected canonicalization\".)\n\n## 6. Recommended Actions\n1. **If any labels are dropped**: seed the source pool with that label, OR extend\n   `ANOMALYGEN_SUPPORTED_LABELS` (and the AnomalyGen generator coverage).\n2. **If source pool is missing**: provide `source_pool_csv` to enable the Mining branch.\n   Without it, half of the augmentation pipeline is dark.\n3. **If AnomalyGen subset is empty**: gap analysis only surfaced labels AnomalyGen cannot\n   generate; rely on Mining for this iteration, or extend the AnomalyGen integration.\n4. **If both subsets are empty**: stop the SDA iteration. Nothing downstream can run.\n",[1324],{"type":48,"tag":86,"props":1325,"children":1326},{"__ignoreMap":367},[1327],{"type":54,"value":1322},{"type":48,"tag":167,"props":1329,"children":1330},{},[],{"type":48,"tag":171,"props":1332,"children":1334},{"id":1333},"execution-order",[1335],{"type":54,"value":1336},"Execution Order",{"type":48,"tag":178,"props":1338,"children":1339},{},[1340,1359,1385,1390],{"type":48,"tag":67,"props":1341,"children":1342},{},[1343,1345,1350,1352,1358],{"type":54,"value":1344},"Run ",{"type":48,"tag":86,"props":1346,"children":1348},{"className":1347},[],[1349],{"type":54,"value":1302},{"type":54,"value":1351}," to get the timestamp; create ",{"type":48,"tag":86,"props":1353,"children":1355},{"className":1354},[],[1356],{"type":54,"value":1357},"\u003Coutput_dir>\u002Frouting_results\u002F\u003Ctimestamp>\u002F",{"type":54,"value":279},{"type":48,"tag":67,"props":1360,"children":1361},{},[1362,1364,1370,1371,1376,1378,1383],{"type":54,"value":1363},"Run the Python recipe (Steps 1–4) to produce ",{"type":48,"tag":86,"props":1365,"children":1367},{"className":1366},[],[1368],{"type":54,"value":1369},"mining_gaps.parquet",{"type":54,"value":118},{"type":48,"tag":86,"props":1372,"children":1374},{"className":1373},[],[1375],{"type":54,"value":579},{"type":54,"value":1377},", and ",{"type":48,"tag":86,"props":1379,"children":1381},{"className":1380},[],[1382],{"type":54,"value":672},{"type":54,"value":1384},". Print summary stats to stdout so the script-check hook can verify it ran.",{"type":48,"tag":67,"props":1386,"children":1387},{},[1388],{"type":54,"value":1389},"Build the per-label decision table by reading both parquets and computing the routed-to verdict per label.",{"type":48,"tag":67,"props":1391,"children":1392},{},[1393,1395,1400],{"type":54,"value":1394},"Write ",{"type":48,"tag":86,"props":1396,"children":1398},{"className":1397},[],[1399],{"type":54,"value":1280},{"type":54,"value":1401}," last — writing it triggers the packaging hook, which copies session logs and skill config alongside.",{"type":48,"tag":1403,"props":1404,"children":1405},"style",{},[1406],{"type":54,"value":1407},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1409,"total":1565},[1410,1428,1445,1456,1468,1482,1495,1509,1522,1533,1547,1556],{"slug":1411,"name":1411,"fn":1412,"description":1413,"org":1414,"tags":1415,"stars":1425,"repoUrl":1426,"updatedAt":1427},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1416,1419,1422],{"name":1417,"slug":1418,"type":15},"Documentation","documentation",{"name":1420,"slug":1421,"type":15},"MCP","mcp",{"name":1423,"slug":1424,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1429,"name":1429,"fn":1430,"description":1431,"org":1432,"tags":1433,"stars":1442,"repoUrl":1443,"updatedAt":1444},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1434,1437,1440],{"name":1435,"slug":1436,"type":15},"Containers","containers",{"name":1438,"slug":1439,"type":15},"Deployment","deployment",{"name":1441,"slug":366,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1446,"name":1446,"fn":1447,"description":1448,"org":1449,"tags":1450,"stars":1442,"repoUrl":1443,"updatedAt":1455},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1451,1454],{"name":1452,"slug":1453,"type":15},"CI\u002FCD","ci-cd",{"name":1438,"slug":1439,"type":15},"2026-07-14T05:25:59.97109",{"slug":1457,"name":1457,"fn":1458,"description":1459,"org":1460,"tags":1461,"stars":1442,"repoUrl":1443,"updatedAt":1467},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1462,1463,1464],{"name":1452,"slug":1453,"type":15},{"name":1438,"slug":1439,"type":15},{"name":1465,"slug":1466,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1469,"name":1469,"fn":1470,"description":1471,"org":1472,"tags":1473,"stars":1442,"repoUrl":1443,"updatedAt":1481},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1474,1477,1478],{"name":1475,"slug":1476,"type":15},"Debugging","debugging",{"name":1465,"slug":1466,"type":15},{"name":1479,"slug":1480,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1483,"name":1483,"fn":1484,"description":1485,"org":1486,"tags":1487,"stars":1442,"repoUrl":1443,"updatedAt":1494},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1488,1491],{"name":1489,"slug":1490,"type":15},"Best Practices","best-practices",{"name":1492,"slug":1493,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1496,"name":1496,"fn":1497,"description":1498,"org":1499,"tags":1500,"stars":1442,"repoUrl":1443,"updatedAt":1508},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1501,1504,1507],{"name":1502,"slug":1503,"type":15},"Machine Learning","machine-learning",{"name":1505,"slug":1506,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1510,"name":1510,"fn":1511,"description":1512,"org":1513,"tags":1514,"stars":1442,"repoUrl":1443,"updatedAt":1521},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1515,1518],{"name":1516,"slug":1517,"type":15},"QA","qa",{"name":1519,"slug":1520,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1523,"name":1523,"fn":1524,"description":1525,"org":1526,"tags":1527,"stars":1442,"repoUrl":1443,"updatedAt":1532},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1528,1529],{"name":1438,"slug":1439,"type":15},{"name":1530,"slug":1531,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1534,"name":1534,"fn":1535,"description":1536,"org":1537,"tags":1538,"stars":1442,"repoUrl":1443,"updatedAt":1546},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1539,1542,1543],{"name":1540,"slug":1541,"type":15},"Code Review","code-review",{"name":1465,"slug":1466,"type":15},{"name":1544,"slug":1545,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1548,"name":1548,"fn":1549,"description":1550,"org":1551,"tags":1552,"stars":1442,"repoUrl":1443,"updatedAt":1555},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1553,1554],{"name":1516,"slug":1517,"type":15},{"name":1519,"slug":1520,"type":15},"2026-07-14T05:25:54.928983",{"slug":1557,"name":1557,"fn":1558,"description":1559,"org":1560,"tags":1561,"stars":1442,"repoUrl":1443,"updatedAt":1564},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1562,1563],{"name":17,"slug":18,"type":15},{"name":1452,"slug":1453,"type":15},"2026-07-30T05:29:03.275638",496,{"items":1567,"total":1663},[1568,1585,1595,1609,1619,1634,1649],{"slug":1569,"name":1569,"fn":1570,"description":1571,"org":1572,"tags":1573,"stars":20,"repoUrl":21,"updatedAt":1584},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1574,1577,1580,1581],{"name":1575,"slug":1576,"type":15},"Data Analysis","data-analysis",{"name":1578,"slug":1579,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":1582,"slug":1583,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":1586,"name":1586,"fn":1587,"description":1588,"org":1589,"tags":1590,"stars":20,"repoUrl":21,"updatedAt":1594},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1591,1592,1593],{"name":1438,"slug":1439,"type":15},{"name":1530,"slug":1531,"type":15},{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1596,"name":1596,"fn":1597,"description":1598,"org":1599,"tags":1600,"stars":20,"repoUrl":21,"updatedAt":1608},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1601,1604,1605],{"name":1602,"slug":1603,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1606,"slug":1607,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1610,"name":1610,"fn":1611,"description":1612,"org":1613,"tags":1614,"stars":20,"repoUrl":21,"updatedAt":1618},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1615,1616,1617],{"name":1575,"slug":1576,"type":15},{"name":9,"slug":8,"type":15},{"name":1519,"slug":1520,"type":15},"2026-07-17T05:29:03.913266",{"slug":1620,"name":1620,"fn":1621,"description":1622,"org":1623,"tags":1624,"stars":20,"repoUrl":21,"updatedAt":1633},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1625,1626,1629,1630],{"name":17,"slug":18,"type":15},{"name":1627,"slug":1628,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1631,"slug":1632,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1635,"name":1635,"fn":1636,"description":1637,"org":1638,"tags":1639,"stars":20,"repoUrl":21,"updatedAt":1648},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1640,1641,1644,1645],{"name":1438,"slug":1439,"type":15},{"name":1642,"slug":1643,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1646,"slug":1647,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1650,"name":1650,"fn":1651,"description":1652,"org":1653,"tags":1654,"stars":20,"repoUrl":21,"updatedAt":1662},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1655,1656,1659],{"name":9,"slug":8,"type":15},{"name":1657,"slug":1658,"type":15},"Quantum Computing","quantum-computing",{"name":1660,"slug":1661,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305]