Changeset 819

Timestamp:

10/27/10 08:49:35 (14 years ago)

Author:

vronk

Message:

bigger rework, get rid of the path-based identifiers =
filling ResourceRef? already with handles read from the MdSelfLink? of the individual mdrecords;
also reading ProviderURL-file and filling as ID in the minimal collection-profile

File:

• metadata/trunk/toolkit/scripts/dir2cmdicollection.py (modified) (4 diffs)

metadata/trunk/toolkit/scripts/dir2cmdicollection.py

@@ -3 +3 @@
 # generates CMDI collection file hierarchy for collections of CMDI records
 # support and questions: Dieter Van Uytvanck <dietuyt@mpi.nl>
+# rework by Matej id@vronk.net : 
+#       - already filling ResourceRef with handles read from the MdSelfLink of the mdrecords
+#   - also reading ProviderURL-file and filling as ID in the basic collection-profile
+#   - does NOT add IsPartOf-elements yet
 
 import os, datetime
 from string import Template
 
+target_dir = "_corpusstructure/"
+
 def main():
         rootList = []
+        if not os.path.isdir(target_dir):
+                os.mkdir(target_dir)
         for root, dirs, files in os.walk(os.getcwd()):
                 startpath = os.getcwd()         
@@ -14 +22 @@
                         if d == "0":
                                 rootList.append(generate_branch(root, dirs))
-        writeCollection(rootList, "collection_root.cmdi")
+        writeCollection(rootList, target_dir + "collection_root.cmdi", "olac-root")
                 
 def generate_branch(root, dirs):
-        collectionFile = "collection_%s.cmdi" % os.path.relpath(root)
+        collectionName = os.path.relpath(root)
+        collectionFile = "_corpusstructure/collection_%s.cmdi" % collectionName
+        
         dirs.sort()
         collectionList = []     
@@ -26 +36 @@
                                 newFile = os.path.relpath(os.path.join(fullpath,file))
                                 collectionList.append(newFile)          
-        writeCollection(collectionList, collectionFile)
-        return collectionFile
+        collid = writeCollection(collectionList, collectionFile, collectionName)
+        print "genbranch:" + collid
+        return collid
 
 
-def writeCollection(collectionList, collectionFile):
+def writeCollection(collectionList, collectionFile, collectionName):
 
-        resourceTemplate = Template('<ResourceProxy id="$idname"><ResourceType>Metadata</ResourceType><ResourceRef>$filename</ResourceRef></ResourceProxy>')
+        resourceTemplate = Template('<ResourceProxy id="$idname"><ResourceType>Metadata</ResourceType><ResourceRef>$idx</ResourceRef></ResourceProxy>')
 
         outstring = Template("""<?xml version="1.0" encoding="UTF-8"?>
 <CMD xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
-    xsi:schemaLocation="http://www.clarin.eu/cmd http://catalog.clarin.eu/ds/ComponentRegistry/rest/registry/profiles/clarin.eu:cr1:p_1271859438236/xsd">
+    xsi:schemaLocation="http://www.clarin.eu/cmd http://catalog.clarin.eu/ds/ComponentRegistry/rest/registry/profiles/clarin.eu:cr1:p_1284723009187/xsd">
     <Header>
         <MdCreator>dir2cmdicollection.py</MdCreator>
         <MdCreationDate>$date</MdCreationDate>
         <MdSelfLink>$selflink</MdSelfLink>
-        <MdProfile>clarin.eu:cr1:p_1271859438236</MdProfile>
+        <MdProfile>clarin.eu:cr1:p_1284723009187</MdProfile>
     </Header>
     <Resources>
@@ -48 +59 @@
         <JournalFileProxyList/>
         <ResourceRelationList/>
-    </Resources>
+    </Resources>    
     <Components>
-        <olac></olac>
+        <collection>
+                <GeneralInfo>
+                <Name>$name</Name>
+                        <ID>$url</ID>
+                </GeneralInfo>
+        </collection>
     </Components>
 </CMD>""")
 
         resourceProxies = ""
-        collectionList.sort()   
+        collectionList.sort()           
+        if os.path.isfile(collectionName + "/ProviderURL"):
+                urlf = open(collectionName + "/ProviderURL", 'r')
+                url = urlf.readline()
+        else: 
+          url ="?"
+        name = "OLAC: " + collectionName.replace("_", " ")
+        idx = ""
         for item in collectionList:
-                resourceProxies += "\n" + resourceTemplate.substitute(idname = item.replace("/", "_").replace("\\", "_"), filename = item)
-        outfile = outstring.substitute(date= datetime.datetime.now().strftime("%Y-%m-%d"), selflink=collectionFile, rp=resourceProxies)
-        f = open(collectionFile, 'w')
+                # trying to restore the original id (which is in the MdSelfLink
+                if os.path.isfile(item):
+                                for line in open(item):
+                                        if "<MdSelfLink>" in line:
+                                                #  WARNING! rocket science employed here !
+                                                idx = line.replace("<MdSelfLink>","").replace("</MdSelfLink>","").strip() 
+                                                break
+                else:
+                         idx = item
+                #idx = item.replace(".xml.cmdi","").replace("_", ":",1)[::-1].replace("_", ":",1)[::-1].replace("_", "-")
+                resourceProxies += "\n" + resourceTemplate.substitute(idname = idx.replace(".","_").replace("/","_"), idx = idx)
+        if collectionName=="olac-root":
+                collidx = "olac-root"
+        else:
+                # print "idx:" + idx
+                if idx!="":
+                        collidx = idx[:idx.rfind(":")] # this is just a hack to derive the collection-id from the id of the collection-item (stripping the running number)
+                else:
+                        collidx = "olac:" + collectionName
+        print collidx 
+        outfile = outstring.substitute(date= datetime.datetime.now().strftime("%Y-%m-%d"), selflink=collidx, rp=resourceProxies,url=url, name=name)
+        f = open(collectionFile, 'w')   
         f.write(outfile)
         f.close()
+        
         print collectionFile
+        return collidx
 
 main()