Merge pull request #7 from MrSampson/main

Add support for other languages

Pipfile

@@ -0,0 +1,18 @@
+[[source]]
+url = "https://pypi.org/simple"
+verify_ssl = true
+name = "pypi"
+
+[packages]
+argparse = "==1.4.0"
+sparqlwrapper = "==1.8.5"
+pandas = "==1.2.4"
+rdflib = "==6.0.2"
+requests = "==2.25.1"
+bs4 = "==0.0.1"
+beautifulsoup4 = "==4.9.3"
+
+[dev-packages]
+
+[requires]
+python_version = "3.12"

eurlex/__init__.py

@@ -321,13 +321,15 @@ def simplify_iri(iri: str) -> str:
     return iri
 
 
-def get_html_by_cellar_id(cellar_id: str) -> str:
+def get_html_by_cellar_id(cellar_id: str, language: str = "en") -> str:
     """Retrieve HTML by CELLAR ID.
 
     Parameters
     ----------
     cellar_id : str
         The CELLAR ID to find HTML for.
+    language : str
+        The language to retrieve the HTML in (default: "en").
 
     Returns
     -------
@@ -342,20 +344,23 @@ def get_html_by_cellar_id(cellar_id: str) -> str:
         allow_redirects=True,
         headers={  # pragma: no cover
             "Accept": "text/html,application/xhtml+xml,application/xml",  # pragma: no cover
-            "Accept-Language": "en",  # pragma: no cover
+            "Accept-Language": f"{language}",  # pragma: no cover
         },
     )  # pragma: no cover
     html = response.content.decode("utf-8")  # pragma: no cover
     return html  # pragma: no cover
 
 
-def get_html_by_celex_id(celex_id: str) -> str:
+def get_html_by_celex_id(celex_id: str, language: str = "en") -> str:
     """Retrieve HTML by CELEX ID.
 
     Parameters
     ----------
     celex_id : str
         The CELEX ID to find HTML for.
+    language : str
+        The language to retrieve the HTML in (default: "en").
+
 
     Returns
     -------
@@ -370,7 +375,7 @@ def get_html_by_celex_id(celex_id: str) -> str:
         allow_redirects=True,
         headers={  # pragma: no cover
             "Accept": "text/html,application/xhtml+xml,application/xml",  # pragma: no cover
-            "Accept-Language": "en",  # pragma: no cover
+            "Accept-Language": f"{language}",  # pragma: no cover
         },
     )  # pragma: no cover
     html = response.content.decode("utf-8")  # pragma: no cover