Upload catalog.py with huggingface_hub
Browse files- catalog.py +31 -10
catalog.py
CHANGED
|
@@ -1,6 +1,8 @@
|
|
|
|
|
| 1 |
import os
|
| 2 |
import re
|
| 3 |
from pathlib import Path
|
|
|
|
| 4 |
|
| 5 |
import requests
|
| 6 |
|
|
@@ -34,16 +36,19 @@ class LocalCatalog(Catalog):
|
|
| 34 |
location: str = default_catalog_path
|
| 35 |
|
| 36 |
def path(self, artifact_identifier: str):
|
| 37 |
-
assert
|
|
|
|
|
|
|
| 38 |
parts = artifact_identifier.split(COLLECTION_SEPARATOR)
|
| 39 |
parts[-1] = parts[-1] + ".json"
|
| 40 |
return os.path.join(self.location, *parts)
|
| 41 |
|
| 42 |
def load(self, artifact_identifier: str):
|
| 43 |
-
assert
|
|
|
|
|
|
|
| 44 |
path = self.path(artifact_identifier)
|
| 45 |
-
|
| 46 |
-
return artifact_instance
|
| 47 |
|
| 48 |
def __getitem__(self, name) -> Artifact:
|
| 49 |
return self.load(name)
|
|
@@ -56,8 +61,16 @@ class LocalCatalog(Catalog):
|
|
| 56 |
return False
|
| 57 |
return os.path.exists(path) and os.path.isfile(path)
|
| 58 |
|
| 59 |
-
def save_artifact(
|
| 60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
if not overwrite:
|
| 62 |
assert (
|
| 63 |
artifact_identifier not in self
|
|
@@ -65,7 +78,8 @@ class LocalCatalog(Catalog):
|
|
| 65 |
path = self.path(artifact_identifier)
|
| 66 |
os.makedirs(Path(path).parent.absolute(), exist_ok=True)
|
| 67 |
artifact.save(path)
|
| 68 |
-
|
|
|
|
| 69 |
|
| 70 |
|
| 71 |
class EnvironmentLocalCatalog(LocalCatalog):
|
|
@@ -97,16 +111,23 @@ class GithubCatalog(LocalCatalog):
|
|
| 97 |
def verify_legal_catalog_name(name):
|
| 98 |
assert re.match(
|
| 99 |
r"^[\w" + COLLECTION_SEPARATOR + "]+$", name
|
| 100 |
-
), '
|
| 101 |
|
| 102 |
|
| 103 |
def add_to_catalog(
|
| 104 |
-
artifact: Artifact,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
):
|
| 106 |
if catalog is None:
|
| 107 |
if catalog_path is None:
|
| 108 |
catalog_path = default_catalog_path
|
| 109 |
catalog = LocalCatalog(location=catalog_path)
|
| 110 |
verify_legal_catalog_name(name)
|
| 111 |
-
catalog.save_artifact(
|
|
|
|
|
|
|
| 112 |
# verify name
|
|
|
|
| 1 |
+
import logging
|
| 2 |
import os
|
| 3 |
import re
|
| 4 |
from pathlib import Path
|
| 5 |
+
from typing import Optional
|
| 6 |
|
| 7 |
import requests
|
| 8 |
|
|
|
|
| 36 |
location: str = default_catalog_path
|
| 37 |
|
| 38 |
def path(self, artifact_identifier: str):
|
| 39 |
+
assert (
|
| 40 |
+
artifact_identifier.strip()
|
| 41 |
+
), "artifact_identifier should not be an empty string."
|
| 42 |
parts = artifact_identifier.split(COLLECTION_SEPARATOR)
|
| 43 |
parts[-1] = parts[-1] + ".json"
|
| 44 |
return os.path.join(self.location, *parts)
|
| 45 |
|
| 46 |
def load(self, artifact_identifier: str):
|
| 47 |
+
assert (
|
| 48 |
+
artifact_identifier in self
|
| 49 |
+
), f"Artifact with name {artifact_identifier} does not exist"
|
| 50 |
path = self.path(artifact_identifier)
|
| 51 |
+
return Artifact.load(path)
|
|
|
|
| 52 |
|
| 53 |
def __getitem__(self, name) -> Artifact:
|
| 54 |
return self.load(name)
|
|
|
|
| 61 |
return False
|
| 62 |
return os.path.exists(path) and os.path.isfile(path)
|
| 63 |
|
| 64 |
+
def save_artifact(
|
| 65 |
+
self,
|
| 66 |
+
artifact: Artifact,
|
| 67 |
+
artifact_identifier: str,
|
| 68 |
+
overwrite: bool = False,
|
| 69 |
+
verbose: bool = True,
|
| 70 |
+
):
|
| 71 |
+
assert isinstance(
|
| 72 |
+
artifact, Artifact
|
| 73 |
+
), f"Input artifact must be an instance of Artifact, got {type(artifact)}"
|
| 74 |
if not overwrite:
|
| 75 |
assert (
|
| 76 |
artifact_identifier not in self
|
|
|
|
| 78 |
path = self.path(artifact_identifier)
|
| 79 |
os.makedirs(Path(path).parent.absolute(), exist_ok=True)
|
| 80 |
artifact.save(path)
|
| 81 |
+
if verbose:
|
| 82 |
+
logging.info(f"Artifact {artifact_identifier} saved to {path}")
|
| 83 |
|
| 84 |
|
| 85 |
class EnvironmentLocalCatalog(LocalCatalog):
|
|
|
|
| 111 |
def verify_legal_catalog_name(name):
|
| 112 |
assert re.match(
|
| 113 |
r"^[\w" + COLLECTION_SEPARATOR + "]+$", name
|
| 114 |
+
), f'Artifict name ("{name}") should be alphanumeric. Use "." for nesting (e.g. myfolder.my_artifact)'
|
| 115 |
|
| 116 |
|
| 117 |
def add_to_catalog(
|
| 118 |
+
artifact: Artifact,
|
| 119 |
+
name: str,
|
| 120 |
+
catalog: Catalog = None,
|
| 121 |
+
overwrite: bool = False,
|
| 122 |
+
catalog_path: Optional[str] = None,
|
| 123 |
+
verbose=True,
|
| 124 |
):
|
| 125 |
if catalog is None:
|
| 126 |
if catalog_path is None:
|
| 127 |
catalog_path = default_catalog_path
|
| 128 |
catalog = LocalCatalog(location=catalog_path)
|
| 129 |
verify_legal_catalog_name(name)
|
| 130 |
+
catalog.save_artifact(
|
| 131 |
+
artifact, name, overwrite=overwrite, verbose=verbose
|
| 132 |
+
) # remove collection (its actually the dir).
|
| 133 |
# verify name
|