From 0af92f46af893b0718f9a400ded11d012814bfcc Mon Sep 17 00:00:00 2001
From: cboulanger <info@bibliograph.org>
Date: Sat, 3 Aug 2024 22:36:58 +0200
Subject: [PATCH] Trying to make the XSLTs work, no luck
---
convert-anystyle-data/anystyle-to-tei.ipynb | 143 ++++++++++++--------
1 file changed, 86 insertions(+), 57 deletions(-)
diff --git a/convert-anystyle-data/anystyle-to-tei.ipynb b/convert-anystyle-data/anystyle-to-tei.ipynb
index 395c904..b9994e2 100644
--- a/convert-anystyle-data/anystyle-to-tei.ipynb
+++ b/convert-anystyle-data/anystyle-to-tei.ipynb
@@ -418,82 +418,111 @@
"metadata": {},
"cell_type": "markdown",
"source": [
- "## Extract bibliographic data from TEI files "
+ "## Extract bibliographic data from TEI files using XSLT\n",
+ "\n",
+ "https://github.com/OpenArabicPE/convert_tei-to-bibliographic-data"
],
"id": "b0a231dc7bdd8b01"
},
- {
- "cell_type": "markdown",
- "source": [
- "### Download XSLTs"
- ],
- "metadata": {
- "collapsed": false
- },
- "id": "149588c08747c4b3"
- },
{
"cell_type": "code",
- "execution_count": 6,
- "outputs": [],
"source": [
- "import requests, zipfile, io, os\n",
- "\n",
- "if not os.path.isdir('lib/convert'): \n",
- " url = 'https://github.com/OpenArabicPE/convert_tei-to-bibliographic-data/archive/refs/heads/master.zip'\n",
- " r = requests.get(url)\n",
- " assert r.status_code == 200\n",
- " z = zipfile.ZipFile(io.BytesIO(r.content))\n",
- " z.extractall('lib')\n",
- " z.close()\n",
- " os.rename('lib/convert_tei-to-bibliographic-data-master', 'lib/convert')\n"
+ "from lxml import etree\n",
+ "import glob\n",
+ "from urllib.request import urlopen\n",
+ "import requests\n",
+ "import traceback\n",
+ "\n",
+ "class HttpsResolver(etree.Resolver):\n",
+ " def resolve(self, url, id, context): \n",
+ " r = requests.get(url)\n",
+ " assert(r.status_code == 200)\n",
+ " return self.resolve_string(r.content, context, base_url=url)\n",
+ "\n",
+ "def apply_xslt(xslt_path, xml_input_path, xml_output_path):\n",
+ " try:\n",
+ " if xslt_path.startswith('http'):\n",
+ " with urlopen(xslt_path) as f:\n",
+ " xml_parser = etree.XMLParser(no_network=False)\n",
+ " xml_parser.resolvers.add(HttpsResolver())\n",
+ " xslt_doc = etree.parse(f, parser=xml_parser)\n",
+ " else:\n",
+ " xslt_doc = etree.parse(xslt_path)\n",
+ " xml_doc = etree.parse(xml_input_path)\n",
+ " transformer = etree.XSLT(xslt_doc)\n",
+ " new_xml = transformer(xml_doc)\n",
+ " with open(xml_output_path, 'w', encoding='utf-8') as f:\n",
+ " f.write(new_xml)\n",
+ " except etree.XSLTParseError as e:\n",
+ " print(f\"Error parsing XSLT file at {xslt_path}: {e}\")\n",
+ "\n",
+ "xslt_url = 'https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl'\n",
+ "\n",
+ "for input_path in glob.glob('tei/*.xml'):\n",
+ " print(f'Converting {input_path}')\n",
+ " base_name = os.path.basename(input_path)\n",
+ " output_path = f'tei-biblstruct/{base_name}'\n",
+ " apply_xslt(xslt_url, input_path, output_path )\n"
],
"metadata": {
"collapsed": false,
"ExecuteTime": {
- "end_time": "2024-07-31T19:42:48.723119300Z",
- "start_time": "2024-07-31T19:42:47.234795500Z"
+ "end_time": "2024-08-03T20:28:20.087934Z",
+ "start_time": "2024-08-03T20:28:18.699591Z"
}
},
- "id": "1f15b3af6aab73ed"
- },
- {
- "metadata": {},
- "cell_type": "markdown",
- "source": [
- "### Apply XSLT"
+ "id": "cb3b4140ab153c08",
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Converting tei/10.1515_zfrs-1980-0103.xml\n",
+ "Error parsing XSLT file at https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl: Failed to compile predicate\n",
+ "Converting tei/10.1515_zfrs-1980-0104.xml\n",
+ "Error parsing XSLT file at https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl: Failed to compile predicate\n",
+ "Converting tei/10.1111_1467-6478.00080.xml\n",
+ "Error parsing XSLT file at https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl: Failed to compile predicate\n",
+ "Converting tei/10.1111_1467-6478.00057.xml\n",
+ "Error parsing XSLT file at https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl: Failed to compile predicate\n"
+ ]
+ }
],
- "id": "aa86435960e61937"
+ "execution_count": 28
},
{
- "cell_type": "code",
- "execution_count": null,
- "outputs": [],
- "source": [
- "from lxml import etree\n",
- "\n",
- "def apply_xslt_to_xml(xslt_path, xml_path):\n",
- " xslt = etree.parse(xslt_path)\n",
- " xml = etree.parse(xml_path)\n",
- " transformer = etree.XSLT(xslt)\n",
- " new_xml = transformer(xml)\n",
- " return str(new_xml)\n",
- "\n",
- "new_xml_str = apply_xslt_to_xml('path_to_your_xslt_file', 'path_to_your_xml_file')\n",
- "print(new_xml_str)\n"
- ],
"metadata": {
- "collapsed": false
+ "ExecuteTime": {
+ "end_time": "2024-08-03T20:28:45.266893Z",
+ "start_time": "2024-08-03T20:28:42.357601Z"
+ }
},
- "id": "cb3b4140ab153c08"
+ "cell_type": "code",
+ "source": "!saxon -s:\"tei/10.1111_1467-6478.00057.xml\" -xsl:\"https://openarabicpe.github.io/convert_tei-to-bibliographic-data/xslt/convert_tei-to-biblstruct_bibl.xsl\"",
+ "id": "2e6d27dc670c0038",
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Error on line 6 column 88 of functions.xsl:\r\n",
+ " XTSE0165 I/O error reported by XML parser processing\r\n",
+ " https://openarabicpe.github.io/../xslt-calendar-conversion/functions/date-functions.xsl.\r\n",
+ " Caused by java.io.IOException: Server returned HTTP response code: 400 for URL:\r\n",
+ " https://openarabicpe.github.io/../xslt-calendar-conversion/functions/date-functions.xsl\r\n",
+ "I/O error reported by XML parser processing https://openarabicpe.github.io/../xslt-calendar-conversion/functions/date-functions.xsl\r\n"
+ ]
+ }
+ ],
+ "execution_count": 29
},
{
- "cell_type": "markdown",
- "source": [],
- "metadata": {
- "collapsed": false
- },
- "id": "387b5b9792505b13"
+ "metadata": {},
+ "cell_type": "code",
+ "outputs": [],
+ "execution_count": null,
+ "source": "",
+ "id": "1a8a57560f1f4868"
}
],
"metadata": {
--
GitLab