Relationship with youseedy? about youseedee HOT 6 OPEN

Yeah, we talked about this in harfbuzz/youseedy#1 :-)

Hmm. Getting all the properties sounds like a good idea, but... it's very slow to parse the whole file, even with lxml: on my computer, 25 seconds with Unihan, 7 seconds without Unihan, compared to 0.7 seconds to parse all the text files.

But what if we don't read it as XML?

from lxml import etree
import re

f = open("ucd.all.flat.xml")
ucd = {}
for l in f.readlines():
    m = re.search(r'char cp="([0-9A-F]+)"', l)
    if m:
        ucd[int(m[1],16)] = l

def getucd(cp):
    return etree.fromstring(ucd[cp].strip()).attrib

print(getucd(0x600))

1.5 seconds. Not bad.

from youseedee.

And if you get very friendly with the file layout, you don't even need regular expressions:

from lxml import etree

f = open("ucd.all.flat.xml")
ucd = {}
for l in f.readlines():
    if "char cp" in l:
        cp = l[16:l.index(r'" age')] # "{6 spaces}<char cp="XXXX" age="
        ucd[int(cp,16)] = l

def getucd(cp):
    return etree.fromstring(ucd[cp].strip()).attrib

print(getucd(0x600))

from youseedee.

Oof. Interesting. Something to pursue definitely.

from youseedee.

Needs to handle groups as well.

from youseedee.

I mean, the group file is much smaller than the flat file. But maybe not after compression. Have to check.

from youseedee.

I was going to download ucdxml and compare sizes but gave up on the front page this time...
https://twitter.com/behdadesfahbod/status/1312090058548154368

from youseedee.