aboutsummaryrefslogtreecommitdiff
path: root/backend/data/dbpedia/genData.py
diff options
context:
space:
mode:
authorTerry Truong <terry06890@gmail.com>2022-05-18 21:09:05 +1000
committerTerry Truong <terry06890@gmail.com>2022-05-18 21:42:32 +1000
commit52560266b585e63742a81e27a3b6f1ef194470c6 (patch)
treed1dd7bb0b0f778d59bf8ed62a5360daf747a737f /backend/data/dbpedia/genData.py
parenteaaa97c186a2f8e2ba0768bd208120c0054ec0d3 (diff)
Add wikipedia-link to tile-info display
Add 'wiki_id' and 'from_dbp' columns to 'descs' table, adjust dbpedia data to include wikipedia IDs, adjust data generations scripts, make server send extra data, and make TileInfo display it.
Diffstat (limited to 'backend/data/dbpedia/genData.py')
-rwxr-xr-xbackend/data/dbpedia/genData.py25
1 files changed, 24 insertions, 1 deletions
diff --git a/backend/data/dbpedia/genData.py b/backend/data/dbpedia/genData.py
index e147641..3df1961 100755
--- a/backend/data/dbpedia/genData.py
+++ b/backend/data/dbpedia/genData.py
@@ -4,13 +4,14 @@ import sys, re
import bz2, sqlite3
usageInfo = f"usage: {sys.argv[0]}\n"
-usageInfo += "Reads DBpedia labels+types+redirects+abstracts data,\n"
+usageInfo += "Reads DBpedia labels/types/abstracts/etc data,\n"
usageInfo += "and creates a sqlite db containing that data.\n"
if len(sys.argv) > 1:
print(usageInfo, file=sys.stderr)
sys.exit(1)
labelsFile = "labels_lang=en.ttl.bz2" # Has about 16e6 lines
+idsFile = "page_lang=en_ids.ttl.bz2"
redirectsFile = "redirects_lang=en_transitive.ttl.bz2"
disambigFile = "disambiguations_lang=en.ttl.bz2"
typesFile = "instance-types_lang=en_specific.ttl.bz2"
@@ -39,6 +40,28 @@ with bz2.open(labelsFile, mode='rt') as file:
else:
dbCur.execute("INSERT INTO labels VALUES (?, ?)", (match.group(1), match.group(2)))
dbCon.commit()
+# Read/store wiki page ids
+print("Reading/storing wiki page ids")
+dbCur.execute("CREATE TABLE ids (iri TEXT PRIMARY KEY, id INT)")
+idLineRegex = re.compile(r'<([^>]+)> <[^>]+> "(\d+)".*\n')
+lineNum = 0
+with bz2.open(idsFile, mode='rt') as file:
+ for line in file:
+ lineNum += 1
+ if lineNum % 1e5 == 0:
+ print("Processing line {}".format(lineNum))
+ #
+ match = idLineRegex.fullmatch(line)
+ if match == None:
+ print("ERROR: Line {} has unexpected format".format(lineNum), file=sys.stderr)
+ sys.exit(1)
+ else:
+ try:
+ dbCur.execute("INSERT INTO ids VALUES (?, ?)", (match.group(1), int(match.group(2))))
+ except sqlite3.IntegrityError as e:
+ # Accounts for certain lines that have the same IRI
+ print("Failed to add entry with IRI \"{}\": {}".format(match.group(1), e))
+dbCon.commit()
# Read/store redirects
print("Reading/storing redirection data")
dbCur.execute("CREATE TABLE redirects (iri TEXT PRIMARY KEY, target TEXT)")