Upload config.py with huggingface_hub
Browse files
config.py
CHANGED
|
@@ -3,6 +3,7 @@ from __future__ import annotations
|
|
| 3 |
import csv
|
| 4 |
import os
|
| 5 |
from dataclasses import dataclass
|
|
|
|
| 6 |
from pathlib import Path
|
| 7 |
|
| 8 |
PROJECT_VERSION = "0.6.0-phase4"
|
|
@@ -96,6 +97,63 @@ IAB_RETRIEVAL_PREFIX_CONFIDENCE_THRESHOLDS = {
|
|
| 96 |
}
|
| 97 |
IAB_PARENT_FALLBACK_CONFIDENCE_FLOOR = 0.3
|
| 98 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 99 |
INTENT_TYPE_LABELS = (
|
| 100 |
"informational",
|
| 101 |
"exploratory",
|
|
|
|
| 3 |
import csv
|
| 4 |
import os
|
| 5 |
from dataclasses import dataclass
|
| 6 |
+
from functools import lru_cache
|
| 7 |
from pathlib import Path
|
| 8 |
|
| 9 |
PROJECT_VERSION = "0.6.0-phase4"
|
|
|
|
| 97 |
}
|
| 98 |
IAB_PARENT_FALLBACK_CONFIDENCE_FLOOR = 0.3
|
| 99 |
|
| 100 |
+
|
| 101 |
+
_DEFAULT_MODEL_REPO_ID = "admesh/agentic-intent-classifier"
|
| 102 |
+
|
| 103 |
+
|
| 104 |
+
def _hf_repo_id() -> str:
|
| 105 |
+
return os.environ.get("ADMESH_MODEL_REPO_ID", _DEFAULT_MODEL_REPO_ID).strip() or _DEFAULT_MODEL_REPO_ID
|
| 106 |
+
|
| 107 |
+
|
| 108 |
+
def _hf_revision() -> str | None:
|
| 109 |
+
rev = os.environ.get("ADMESH_MODEL_REVISION", "").strip()
|
| 110 |
+
return rev or None
|
| 111 |
+
|
| 112 |
+
|
| 113 |
+
@lru_cache(maxsize=8)
|
| 114 |
+
def _resolve_repo_subdir(local_dir: Path, repo_subdir: str) -> Path:
|
| 115 |
+
"""Resolve artifact/model subdirs for local dev and HF trust_remote_code.
|
| 116 |
+
|
| 117 |
+
Local runs: return on-disk folder inside repo.
|
| 118 |
+
HF dynamic module runs: if missing locally, pull only this subdir from Hub.
|
| 119 |
+
"""
|
| 120 |
+
if local_dir.exists():
|
| 121 |
+
return local_dir
|
| 122 |
+
|
| 123 |
+
try:
|
| 124 |
+
from huggingface_hub import snapshot_download
|
| 125 |
+
except Exception:
|
| 126 |
+
return local_dir
|
| 127 |
+
|
| 128 |
+
kwargs: dict = {
|
| 129 |
+
"repo_id": _hf_repo_id(),
|
| 130 |
+
"repo_type": "model",
|
| 131 |
+
"allow_patterns": [f"{repo_subdir}/**"],
|
| 132 |
+
}
|
| 133 |
+
revision = _hf_revision()
|
| 134 |
+
if revision:
|
| 135 |
+
kwargs["revision"] = revision
|
| 136 |
+
|
| 137 |
+
try:
|
| 138 |
+
root = Path(snapshot_download(**kwargs))
|
| 139 |
+
candidate = root / repo_subdir
|
| 140 |
+
if candidate.exists():
|
| 141 |
+
return candidate
|
| 142 |
+
except Exception:
|
| 143 |
+
pass
|
| 144 |
+
return local_dir
|
| 145 |
+
|
| 146 |
+
|
| 147 |
+
# Re-resolve critical artifact/model dirs after helper definitions.
|
| 148 |
+
CALIBRATION_ARTIFACTS_DIR = _resolve_repo_subdir(ARTIFACTS_DIR / "calibration", "artifacts/calibration")
|
| 149 |
+
IAB_ARTIFACTS_DIR = _resolve_repo_subdir(ARTIFACTS_DIR / "iab", "artifacts/iab")
|
| 150 |
+
IAB_TAXONOMY_GRAPH_PATH = IAB_ARTIFACTS_DIR / "taxonomy_graph.json"
|
| 151 |
+
IAB_TAXONOMY_NODES_PATH = IAB_ARTIFACTS_DIR / "taxonomy_nodes.json"
|
| 152 |
+
IAB_TAXONOMY_EMBEDDINGS_PATH = IAB_ARTIFACTS_DIR / "taxonomy_embeddings.pt"
|
| 153 |
+
IAB_DATASET_SUMMARY_PATH = IAB_ARTIFACTS_DIR / "dataset_summary.json"
|
| 154 |
+
MULTITASK_INTENT_MODEL_DIR = _resolve_repo_subdir(BASE_DIR / "multitask_intent_model_output", "multitask_intent_model_output")
|
| 155 |
+
IAB_CLASSIFIER_MODEL_DIR = _resolve_repo_subdir(BASE_DIR / "iab_classifier_model_output", "iab_classifier_model_output")
|
| 156 |
+
|
| 157 |
INTENT_TYPE_LABELS = (
|
| 158 |
"informational",
|
| 159 |
"exploratory",
|