dalia_dif 0.0.17__py3-none-any.whl → 0.0.18__py3-none-any.whl

dalia_dif/cli.py

@@ -18,10 +18,19 @@ def main() -> None:
 @main.command()
 @click.option("--dif-version", type=click.Choice(["1.3"]), default="1.3")
 @click.option("--ignore-missing-description", is_flag=True)
+@click.option("--communities-path", type=Path)
 @click.argument("location")
-def validate(location: str, dif_version: str, ignore_missing_description: bool) -> None:
+def validate(
+    location: str, dif_version: str, ignore_missing_description: bool, communities_path: Path | None
+) -> None:
     """Validate a local/remote file or local folder of DIF-encoded CSVs."""
     from dalia_dif.dif13 import read_dif13
+    from dalia_dif.dif13.community import get_communities_dict
+
+    if communities_path is not None:
+        community_dict = get_communities_dict(communities_path)
+    else:
+        community_dict = {}
 
     fail = False
     p = Path(location)
@@ -36,6 +45,7 @@ def validate(location: str, dif_version: str, ignore_missing_description: bool)
                 path,
                 error_accumulator=errors,
                 ignore_missing_description=ignore_missing_description,
+                custom_community_dict=community_dict,
             )
             if errors:
                 fail = True

dalia_dif/dif13/community/__init__.py

@@ -3,35 +3,69 @@
 import csv
 from collections import Counter
 from pathlib import Path
+from typing import Any, TypeAlias
+
+from pydantic import UUID4, AnyHttpUrl, BaseModel, Field
 
 HERE = Path(__file__).parent.resolve()
 COMMUNITIES_PATH = HERE / "dalia_communities.csv"
 
 
-def _read_mapping() -> dict[str, str]:
+class Community(BaseModel):
+    """A data model for communities."""
+
+    uuid: UUID4
+    title: str
+    ror: str | None = None
+    website: AnyHttpUrl | None = None
+    synonyms: list[str] = Field(default_factory=list)
+
+
+def _process(row: dict[str, Any]) -> Community:
+    if synonyms_raw := row.pop("synonyms"):
+        row["synonyms"] = [s.strip() for s in synonyms_raw.split("|")]
+    return Community.model_validate({k: v for k, v in row.items() if v})
+
+
+def read_communities(path: Path) -> list[Community]:
+    """Read communities."""
+    with open(path, newline="") as csvfile:
+        return [_process(row) for row in csv.DictReader(csvfile)]
+
+
+CommunityDict: TypeAlias = dict[str, str]
+
+
+def get_communities_dict(path: Path) -> CommunityDict:
+    """Get a mapping from names/synonyms to UUID strings."""
     rv = {}
-    with open(COMMUNITIES_PATH, newline="") as csvfile:
-        for row in csv.DictReader(csvfile):
-            rv[row["Title"]] = row["ID"]
-            for synonym in row["Synonyms"].split("|"):
-                rv[synonym] = row["ID"]
+    for community in read_communities(path):
+        rv[community.title] = str(community.uuid)
+        for synonym in community.synonyms:
+            rv[synonym] = str(community.uuid)
     return rv
 
 
-LOOKUP_DICT_COMMUNITIES = _read_mapping()
+def _read_mapping() -> dict[str, str]:
+    rv = {}
+    for community in read_communities(COMMUNITIES_PATH):
+        rv[community.title] = str(community.uuid)
+        for synonym in community.synonyms:
+            rv[synonym] = str(community.uuid)
+    return rv
+
 
-del _read_mapping
+LOOKUP_DICT_COMMUNITIES: CommunityDict = get_communities_dict(COMMUNITIES_PATH)
 
 MISSING_COMMUNITIES: Counter[str] = Counter()
 
 
 def get_community_labels() -> dict[str, str]:
     """Get community labels."""
-    with open(COMMUNITIES_PATH, newline="") as csvfile:
-        return {
-            row["ID"]: COMMUNITY_RELABELS.get(row["Title"], row["Title"])
-            for row in csv.DictReader(csvfile)
-        }
+    return {
+        str(community.uuid): COMMUNITY_RELABELS.get(community.title, community.title)
+        for community in read_communities(COMMUNITIES_PATH)
+    }
 
 
 COMMUNITY_RELABELS = {

dalia_dif/dif13/community/dalia_communities.csv

@@ -1,4 +1,4 @@
-ID,ROR,Website,Title,Synonyms
+uuid,ror,website,title,synonyms
 8cd01866-7560-4701-ba4a-da3c939b9061,,,Base4NFDI,
 7ecf1a3e-e377-4f5c-ac70-78d498951843,,,BERD@NFDI,
 0393d642-340d-4641-8c1f-e9c8b27199bf,,,DAPHNE4NFDI,

dalia_dif/dif13/reader.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 import logging
 import re
-from collections import Counter
+from collections import ChainMap, Counter
 from pathlib import Path
 from typing import TextIO
 
@@ -17,7 +17,7 @@ from pystow.utils import safe_open_dict_reader
 from rdflib import URIRef
 from tqdm import tqdm
 
-from .community import LOOKUP_DICT_COMMUNITIES
+from .community import LOOKUP_DICT_COMMUNITIES, CommunityDict
 from .model import (
     AuthorDIF13,
     EducationalResourceDIF13,
@@ -102,6 +102,7 @@ def read_dif13(
     error_accumulator: list[str] | None = None,
     converter: curies.Converter | None = None,
     ignore_missing_description: bool = False,
+    custom_community_dict: CommunityDict | None = None,
 ) -> list[EducationalResourceDIF13]:
     """Parse DALIA records."""
     if isinstance(path, str) and (path.startswith("http://") or path.startswith("https://")):
@@ -139,6 +140,7 @@ def read_dif13(
                     error_accumulator=error_accumulator,
                     converter=converter,
                     ignore_missing_description=ignore_missing_description,
+                    custom_community_dict=custom_community_dict,
                 )
             )
             is not None
@@ -164,13 +166,18 @@ def parse_dif13_row(  # noqa:C901
     error_accumulator: list[str] | None = None,
     converter: curies.Converter | None = None,
     ignore_missing_description: bool = False,
+    custom_community_dict: CommunityDict | None = None,
 ) -> EducationalResourceDIF13 | None:
     """Convert a row in a DALIA curation file to a resource, or return none if unable."""
     if isinstance(file_name, Path):
         file_name = file_name.name
 
     supporting_communities, recommending_communities = _process_communities(
-        file_name, idx, row, error_accumulator=error_accumulator
+        file_name,
+        idx,
+        row,
+        error_accumulator=error_accumulator,
+        custom_community_dict=custom_community_dict,
     )
 
     external_uris = _pop_split(row, "Link")
@@ -473,8 +480,10 @@ def _process_communities(
     row: dict[str, str],
     *,
     error_accumulator: list[str] | None = None,
+    custom_community_dict: CommunityDict | None = None,
 ) -> tuple[list[URIRef], list[URIRef]]:
     supporting, recommending = [], []
+    community_dict = ChainMap(custom_community_dict or {}, LOOKUP_DICT_COMMUNITIES)
     for community in _pop_split(row, "Community"):
         match = COMMUNITY_RELATION_RE.search(community)
         if not match:
@@ -489,7 +498,7 @@ def _process_communities(
         name = match.group("name").strip()
         relation = match.group("relation")
 
-        community_uuid = LOOKUP_DICT_COMMUNITIES.get(name, None)
+        community_uuid = community_dict.get(name, None)
         if not community_uuid:
             if not MISSING_COMMUNITIES[name]:
                 _log(

dalia_dif/version.py

@@ -12,7 +12,7 @@ __all__ = [
     "get_version",
 ]
 
-VERSION = "0.0.17"
+VERSION = "0.0.18"
 
 
 def get_git_hash() -> str:

{dalia_dif-0.0.17.dist-info → dalia_dif-0.0.18.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dalia_dif
-Version: 0.0.17
+Version: 0.0.18
 Summary: Tools for DALIA's data model for open educational resources
 Keywords: snekpack,cookiecutter,open educational resources,educational resources,training
 Author: Charles Tapley Hoyt
@@ -39,9 +39,9 @@ Maintainer: Charles Tapley Hoyt
 Maintainer-email: Charles Tapley Hoyt <cthoyt@gmail.com>
 Requires-Python: >=3.10
 Project-URL: Bug Tracker, https://github.com/data-literacy-alliance/dalia-dif/issues
-Project-URL: Documentation, https://dalia-dif.readthedocs.io
 Project-URL: Homepage, https://github.com/data-literacy-alliance/dalia-dif
 Project-URL: Repository, https://github.com/data-literacy-alliance/dalia-dif.git
+Project-URL: Documentation, https://dalia-dif.readthedocs.io
 Provides-Extra: export
 Provides-Extra: fti
 Description-Content-Type: text/markdown

{dalia_dif-0.0.17.dist-info → dalia_dif-0.0.18.dist-info}/RECORD

@@ -1,10 +1,10 @@
 dalia_dif/.DS_Store,sha256=MtnD5ThzFbk9R7gZVWNniR_qnre6YMJscVWWkkQW8AM,6148
 dalia_dif/__init__.py,sha256=uyO7jo0cu-99M2a87I7XeOH6c9ZPqRH46YgTdfXLOWg,66
 dalia_dif/__main__.py,sha256=dlPNI11KSS49DY4SoTA1l8gQZr6iR9N1JJsySTjicUg,126
-dalia_dif/cli.py,sha256=zIKecW5aCvUrwyeW6k3ZXbb33Jl3X1d-9tZBsRpBLoU,3163
+dalia_dif/cli.py,sha256=yvV0YeBSJ9cGlmcsuTaVqGw6-zokjLvoQ8xYHjPzoq8,3504
 dalia_dif/dif13/__init__.py,sha256=tMaSW0hhnuO_Mt8CM_d0cTT6j_FSTGFY99crUawWa4w,457
-dalia_dif/dif13/community/__init__.py,sha256=y8RnUSR1-d0oU02ZJA6z1hmDIJlUTJ66Yb51JGcp0JE,1051
-dalia_dif/dif13/community/dalia_communities.csv,sha256=Tcpo310kOtn1zNMKu183GjqcM8XlsubGrLSvbvotJxI,6191
+dalia_dif/dif13/community/__init__.py,sha256=IwHOubp6bc6peW2vVgaYuA7JNJ07py60bUZowKDttvs,2107
+dalia_dif/dif13/community/dalia_communities.csv,sha256=Dg4rGox0ykZ-2HFFgVF8hDZIxcX2VgeKUle2W4PHD0c,6193
 dalia_dif/dif13/constants.py,sha256=NLNutj5AYok_WY1t1iE01dOCKK2eCnkezzN4TUgT-SA,6797
 dalia_dif/dif13/export/__init__.py,sha256=-HXO3x8soeYPFg5vIggJcZihGbyeVn_DO42cqWwNAuA,28
 dalia_dif/dif13/export/charts.py,sha256=8cXrq2UMnQUFi_Bda5L5sbZJRu2DnfUftj8zPe5c8ik,13695
@@ -17,7 +17,7 @@ dalia_dif/dif13/model.py,sha256=GIfzAnT0Blh7Qe-wk3w_lxZld-B4VbfH2ahyaDdpuR4,6935
 dalia_dif/dif13/picklists.py,sha256=FapSFANJZ_u432RtGyOq7v8FpKfR4X9C7n3bAAXbrNI,4933
 dalia_dif/dif13/predicates.py,sha256=rrWVPMnMPvMelS4i2hOEWGhL11t-KR8yRrDArhGvYPE,1700
 dalia_dif/dif13/rdf.py,sha256=ZPdGB1hM2zNuxe6gidNwOn0LJB00Zs-mj2fHOYNm_qU,5908
-dalia_dif/dif13/reader.py,sha256=g7ZtDSrUBhlKNYyvB39auAdWQHndGf6pZGJa3jFpWes,17200
+dalia_dif/dif13/reader.py,sha256=SinQMN_cimwihpsM3kJwom-MD8Om_WPTTi5Bo5EOKxQ,17611
 dalia_dif/dif13/utils.py,sha256=-r5ezhD7AkaJUAsUtZl1sHG_P76fz2Q_ULXgXLyDNII,814
 dalia_dif/namespace/__init__.py,sha256=kJGpIXZzeUdVhM_Pj465XlMkVqMB2Uc9aAZeFKQp5IQ,3105
 dalia_dif/namespace/bibframe_lite_relation.py,sha256=2trFP98s1wCPUfiAjHg-EK09yjWXC6ePLNJIVqgZvd8,255
@@ -31,9 +31,9 @@ dalia_dif/namespace/modalia.py,sha256=pSL-G9ib0E--Jwrbu9PvMMRjMe2FzzxoSjEv-0ank2
 dalia_dif/namespace/rec.py,sha256=c2Ce5IEhnM_BcZJUCXiMQEZ2GLFcC_76hfkKE16-zLc,241
 dalia_dif/py.typed,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 dalia_dif/utils.py,sha256=AA19Bw0vv8-ODZzDy_VR57WwndRccaQIepVS8tvodDo,761
-dalia_dif/version.py,sha256=aAPqI66QtdgdBvrGfuSXNCjIUCx6siw8G6_BPaSwT7w,962
-dalia_dif-0.0.17.dist-info/licenses/LICENSE,sha256=S-DsND478R_VQyGmtXbVYW67fRiJj3QfY8UXIJ4zvLI,1076
-dalia_dif-0.0.17.dist-info/WHEEL,sha256=XjEbIc5-wIORjWaafhI6vBtlxDBp7S9KiujWF1EM7Ak,79
-dalia_dif-0.0.17.dist-info/entry_points.txt,sha256=GmBz8JVyUD6m4v2QIxa7gm3sHIVNjTl5KTPP7L3FCGI,50
-dalia_dif-0.0.17.dist-info/METADATA,sha256=JwAwFe9RTi_fMDtlWApattEpFAihCPtSqcG20tF-rKg,17760
-dalia_dif-0.0.17.dist-info/RECORD,,
+dalia_dif/version.py,sha256=7L891_Jrr5liLFWnZMCce8I-vBl683ocjoLsHlmjYGY,962
+dalia_dif-0.0.18.dist-info/licenses/LICENSE,sha256=S-DsND478R_VQyGmtXbVYW67fRiJj3QfY8UXIJ4zvLI,1076
+dalia_dif-0.0.18.dist-info/WHEEL,sha256=fAguSjoiATBe7TNBkJwOjyL1Tt4wwiaQGtNtjRPNMQA,80
+dalia_dif-0.0.18.dist-info/entry_points.txt,sha256=GmBz8JVyUD6m4v2QIxa7gm3sHIVNjTl5KTPP7L3FCGI,50
+dalia_dif-0.0.18.dist-info/METADATA,sha256=rqYxck75cMrF7PtSj4xMR9NeaPe9YySdUEVrW8Cs4Sw,17760
+dalia_dif-0.0.18.dist-info/RECORD,,

{dalia_dif-0.0.17.dist-info → dalia_dif-0.0.18.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: uv 0.9.25
+Generator: uv 0.9.28
 Root-Is-Purelib: true
-Tag: py3-none-any
+Tag: py3-none-any