support latest smileys

This commit is contained in:
Yuxin Wu
2020-06-27 17:31:35 -07:00
parent 628bc024c8
commit 1395351755
34 changed files with 381 additions and 68 deletions
+1 -1
View File
@@ -171,7 +171,7 @@ class HTMLRender(object):
def _render_partial_msgs(self, msgs):
""" return single html"""
self.smiley.used_smiley_id.clear()
self.smiley.reset()
slicer = MessageSlicerByTime()
slices = slicer.slice(msgs)
+25 -55
View File
@@ -2,6 +2,7 @@
# -*- coding: UTF-8 -*-
import os
import functools
import re
import json
import struct
@@ -9,10 +10,7 @@ import struct
from common.textutil import get_file_b64
STATIC_PATH = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'static')
UNICODE_SMILEY_FILE = os.path.join(STATIC_PATH, 'unicode-smiley.json')
TENCENT_SMILEY_FILE = os.path.join(STATIC_PATH, 'tencent-smiley.json')
TENCENT_EXTRASMILEY_FILE = os.path.join(STATIC_PATH, 'tencent-smiley-extra.json')
try:
UNICODE_SMILEY_RE = re.compile(
@@ -37,10 +35,16 @@ HEAD = """.smiley {
}
"""
TEMPLATE = """.smiley{name} {{
TEMPLATE = """.{name} {{
background-image: url("data:image/png;base64,{b64}");
background-size: 24px 24px;
}}"""
def _css_class_name(s):
s = s.replace("/", "_")
s = s.replace(".", "_")
return "smiley_" + s
class SmileyProvider(object):
def __init__(self, html_replace=True):
""" html_replace: replace smileycode by html.
@@ -50,74 +54,40 @@ class SmileyProvider(object):
if not html_replace:
raise NotImplementedError()
# [微笑] -> 0
# [微笑] -> smiley/0.png
self.tencent_smiley = json.load(open(TENCENT_SMILEY_FILE))
self.used_smileys = set()
# some extra smiley from javascript on wx.qq.com
extra_smiley = json.load(open(TENCENT_EXTRASMILEY_FILE))
extra_smiley = {u'[' + k + u']': v for k, v in
extra_smiley.items()}
self.tencent_smiley.update(extra_smiley)
def reset(self):
self.used_smileys.clear()
# 1f35c -> "\ue340"
#self.unicode_smiley_code = gUnicodeCodeMap
# u'\U0001f35c' -> "e340" # for iphone
# u'\ue415' -> 'e415' # for android
unicode_smiley_dict = json.load(open(UNICODE_SMILEY_FILE))
self.unicode_smiley = {(self.unichar(int(k, 16))): hex(ord(v))[2:] for k, v in
unicode_smiley_dict.items()}
self.unicode_smiley.update({v: hex(ord(v))[2:] for _, v in
unicode_smiley_dict.items()})
self.used_smiley_id = set()
def unichar(self, i):
try:
return chr(i)
except ValueError:
return struct.pack('i', i).decode('utf-32')
def gen_replace_elem(self, smiley_id):
self.used_smiley_id.add(str(smiley_id))
return '<span class="smiley smiley{}"></span>'.format(smiley_id)
def _replace_unicode(self, msg):
if not UNICODE_SMILEY_RE.findall(msg):
# didn't find the code
return msg
for k, v in self.unicode_smiley.items():
if k in msg:
msg = msg.replace(k, self.gen_replace_elem(v))
return msg
def _replace_tencent(self, msg):
if (not '[' in msg or not ']' in msg) \
and (not '/:' in msg) and (not '/' in msg):
return msg
for k, v in self.tencent_smiley.items():
if k in msg:
msg = msg.replace(k, self.gen_replace_elem(v))
return msg
def gen_replace_elem(self, smiley_path):
self.used_smileys.add(str(smiley_path))
return '<span class="smiley {}"></span>'.format(_css_class_name(smiley_path))
def replace_smileycode(self, msg):
""" replace the smiley code in msg
return a html
"""
msg = self._replace_unicode(msg)
msg = self._replace_tencent(msg)
# pre-filter:
if ('[' not in msg) and ('/' not in msg) and not UNICODE_SMILEY_RE.findall(msg):
return msg
for k, v in self.tencent_smiley.items():
if k in msg:
msg = msg.replace(k, self.gen_replace_elem(v))
return msg
return msg
def gen_used_smiley_css(self):
ret = HEAD
for sid in self.used_smiley_id:
fname = os.path.join(STATIC_PATH, 'smileys', '{}.png'.format(sid))
for path in self.used_smileys:
fname = os.path.join(STATIC_PATH, path)
b64 = get_file_b64(fname)
ret = ret + TEMPLATE.format(name=sid, b64=b64)
ret = ret + TEMPLATE.format(name=_css_class_name(path), b64=b64)
return ret
if __name__ == '__main__':
smiley = SmileyProvider()
msg = u"[挥手]哈哈呵呵hihi\U0001f684\u2728\u0001 /::<\ue415"
msg = smiley.replace_smileycode(msg)
#print msg
smiley.gen_used_smiley_css()
+10
View File
@@ -0,0 +1,10 @@
Sources:
* web wechat
* /data/data/com.tencent.mm/files/public/emoji/newemoji
Generated file:
* tencent-smiley.json generated by `parse_tencent_smiley.py`
Binary file not shown.

After

Width:  |  Height:  |  Size: 2.2 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.7 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 742 B

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.2 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 890 B

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.9 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 11 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 8.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 13 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 13 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 13 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 14 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 10 KiB

+2
View File
@@ -0,0 +1,2 @@
{
"version": 86}
+275
View File
@@ -0,0 +1,275 @@
<?xml version="1.0" encoding="utf-8"?>
<newemoji>
<emoji>
<key><![CDATA[[Smirk]]]></key>
<cn-value><![CDATA[[奸笑]]]></cn-value>
<qq-value><![CDATA[[奸笑]]]></qq-value>
<en-value><![CDATA[[Smirk]]]></en-value>
<tw-value><![CDATA[[奸笑]]]></tw-value>
<th-value><![CDATA[[Smirk]]]></th-value>
<fileName>2_02.png</fileName>
<eggIndex>2002</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Hey]]]></key>
<cn-value><![CDATA[[嘿哈]]]></cn-value>
<qq-value><![CDATA[[嘿哈]]]></qq-value>
<en-value><![CDATA[[Hey]]]></en-value>
<tw-value><![CDATA[[吼嘿]]]></tw-value>
<th-value><![CDATA[[Hey]]]></th-value>
<fileName>2_04.png</fileName>
<eggIndex>2004</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Facepalm]]]></key>
<cn-value><![CDATA[[捂脸]]]></cn-value>
<qq-value><![CDATA[[捂脸]]]></qq-value>
<en-value><![CDATA[[Facepalm]]]></en-value>
<tw-value><![CDATA[[掩面]]]></tw-value>
<th-value><![CDATA[[Facepalm]]]></th-value>
<fileName>2_05.png</fileName>
<eggIndex>2005</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Smart]]]></key>
<cn-value><![CDATA[[机智]]]></cn-value>
<qq-value><![CDATA[[机智]]]></qq-value>
<en-value><![CDATA[[Smart]]]></en-value>
<tw-value><![CDATA[[機智]]]></tw-value>
<th-value><![CDATA[[Smart]]]></th-value>
<fileName>2_06.png</fileName>
<eggIndex>2006</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Tea]]]></key>
<cn-value><![CDATA[[茶]]]></cn-value>
<qq-value><![CDATA[[茶]]]></qq-value>
<en-value><![CDATA[[Tea]]]></en-value>
<tw-value><![CDATA[[茶]]]></tw-value>
<th-value><![CDATA[[Tea]]]></th-value>
<fileName>2_07.png</fileName>
<eggIndex>2007</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Packet]]]></key>
<cn-value><![CDATA[[红包]]]></cn-value>
<qq-value><![CDATA[[红包]]]></qq-value>
<en-value><![CDATA[[Packet]]]></en-value>
<tw-value><![CDATA[[Packet]]]></tw-value>
<th-value><![CDATA[[Packet]]]></th-value>
<fileName>2_09.png</fileName>
<eggIndex>2009</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Candle]]]></key>
<cn-value><![CDATA[[蜡烛]]]></cn-value>
<qq-value><![CDATA[[蜡烛]]]></qq-value>
<en-value><![CDATA[[Candle]]]></en-value>
<tw-value><![CDATA[[蠟燭]]]></tw-value>
<th-value><![CDATA[[Candle]]]></th-value>
<fileName>2_10.png</fileName>
<eggIndex>2010</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Yeah!]]]></key>
<cn-value><![CDATA[[耶]]]></cn-value>
<qq-value><![CDATA[[耶]]]></qq-value>
<en-value><![CDATA[[Yeah!]]]></en-value>
<tw-value><![CDATA[[歐耶]]]></tw-value>
<th-value><![CDATA[[Yeah!]]]></th-value>
<fileName>2_11.png</fileName>
<eggIndex>2011</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Concerned]]]></key>
<cn-value><![CDATA[[皱眉]]]></cn-value>
<qq-value><![CDATA[[皱眉]]]></qq-value>
<en-value><![CDATA[[Concerned]]]></en-value>
<tw-value><![CDATA[[皺眉]]]></tw-value>
<th-value><![CDATA[[Concerned]]]></th-value>
<fileName>2_12.png</fileName>
<eggIndex>2018</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Blush]]]></key>
<cn-value><![CDATA[[囧]]]></cn-value>
<qq-value><![CDATA[[囧]]]></qq-value>
<en-value><![CDATA[[Blush]]]></en-value>
<tw-value><![CDATA[[囧]]]></tw-value>
<th-value><![CDATA[[Blush]]]></th-value>
<fileName>smiley_17b.png</fileName>
<eggIndex>2013</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Salute]]]></key>
<cn-value><![CDATA[[抱拳]]]></cn-value>
<qq-value><![CDATA[[抱拳]]]></qq-value>
<en-value><![CDATA[[Salute]]]></en-value>
<tw-value><![CDATA[[抱拳]]]></tw-value>
<th-value><![CDATA[[Salute]]]></th-value>
<fileName>smiley_83b.png</fileName>
<eggIndex>2014</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Chick]]]></key>
<cn-value><![CDATA[[鸡]]]></cn-value>
<qq-value><![CDATA[[鸡]]]></qq-value>
<en-value><![CDATA[[Chick]]]></en-value>
<tw-value><![CDATA[[小雞]]]></tw-value>
<th-value><![CDATA[[Chick]]]></th-value>
<fileName>2_14.png</fileName>
<eggIndex>2015</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Blessing]]]></key>
<cn-value><![CDATA[[福]]]></cn-value>
<qq-value><![CDATA[[福]]]></qq-value>
<en-value><![CDATA[[Blessing]]]></en-value>
<tw-value><![CDATA[[福]]]></tw-value>
<th-value><![CDATA[[Blessing]]]></th-value>
<fileName>2_15.png</fileName>
<eggIndex>2016</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Bye]]]></key>
<cn-value><![CDATA[[再见]]]></cn-value>
<qq-value><![CDATA[[再见]]]></qq-value>
<en-value><![CDATA[[Bye]]]></en-value>
<tw-value><![CDATA[[再見]]]></tw-value>
<th-value><![CDATA[[Bye]]]></th-value>
<fileName>smiley_39b.png</fileName>
<eggIndex>2017</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Rich]]]></key>
<cn-value><![CDATA[[發]]]></cn-value>
<qq-value><![CDATA[[發]]]></qq-value>
<en-value><![CDATA[[Rich]]]></en-value>
<tw-value><![CDATA[[發]]]></tw-value>
<th-value><![CDATA[[Rich]]]></th-value>
<fileName>2_16.png</fileName>
<eggIndex>2019</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Pup]]]></key>
<cn-value><![CDATA[[小狗]]]></cn-value>
<qq-value><![CDATA[[小狗]]]></qq-value>
<en-value><![CDATA[[Pup]]]></en-value>
<tw-value><![CDATA[[小狗]]]></tw-value>
<th-value><![CDATA[[Pup]]]></th-value>
<fileName>2_17.png</fileName>
<eggIndex>2020</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Onlooker]]]></key>
<cn-value><![CDATA[[吃瓜]]]></cn-value>
<qq-value><![CDATA[[吃瓜]]]></qq-value>
<en-value><![CDATA[[Onlooker]]]></en-value>
<tw-value><![CDATA[[吃西瓜]]]></tw-value>
<th-value><![CDATA[[Onlooker]]]></th-value>
<fileName>Watermelon.png</fileName>
<eggIndex>2021</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[GoForIt]]]></key>
<cn-value><![CDATA[[加油]]]></cn-value>
<qq-value><![CDATA[[加油]]]></qq-value>
<en-value><![CDATA[[GoForIt]]]></en-value>
<tw-value><![CDATA[[加油]]]></tw-value>
<th-value><![CDATA[[GoForIt]]]></th-value>
<fileName>Addoil.png</fileName>
<eggIndex>2022</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Sweats]]]></key>
<cn-value><![CDATA[[汗]]]></cn-value>
<qq-value><![CDATA[[汗]]]></qq-value>
<en-value><![CDATA[[Sweats]]]></en-value>
<tw-value><![CDATA[[汗]]]></tw-value>
<th-value><![CDATA[[Sweats]]]></th-value>
<fileName>Sweat.png</fileName>
<eggIndex>2023</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[OMG]]]></key>
<cn-value><![CDATA[[天啊]]]></cn-value>
<qq-value><![CDATA[[天啊]]]></qq-value>
<en-value><![CDATA[[OMG]]]></en-value>
<tw-value><![CDATA[[天啊]]]></tw-value>
<th-value><![CDATA[[OMG]]]></th-value>
<fileName>Shocked!.png</fileName>
<eggIndex>2025</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Emm]]]></key>
<cn-value><![CDATA[[Emm]]]></cn-value>
<qq-value><![CDATA[[Emm]]]></qq-value>
<en-value><![CDATA[[Emm]]]></en-value>
<tw-value><![CDATA[[一言難盡]]]></tw-value>
<th-value><![CDATA[[Emm]]]></th-value>
<fileName>Cold.png</fileName>
<eggIndex>2027</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Respect]]]></key>
<cn-value><![CDATA[[社会社会]]]></cn-value>
<qq-value><![CDATA[[社会社会]]]></qq-value>
<en-value><![CDATA[[Respect]]]></en-value>
<tw-value><![CDATA[[失敬失敬]]]></tw-value>
<th-value><![CDATA[[Respect]]]></th-value>
<fileName>Social.png</fileName>
<eggIndex>2028</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Doge]]]></key>
<cn-value><![CDATA[[旺柴]]]></cn-value>
<qq-value><![CDATA[[旺柴]]]></qq-value>
<en-value><![CDATA[[Doge]]]></en-value>
<tw-value><![CDATA[[旺柴]]]></tw-value>
<th-value><![CDATA[[Doge]]]></th-value>
<fileName>Yellowdog.png</fileName>
<eggIndex>2030</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[NoProb]]]></key>
<cn-value><![CDATA[[好的]]]></cn-value>
<qq-value><![CDATA[[好的]]]></qq-value>
<en-value><![CDATA[[NoProb]]]></en-value>
<tw-value><![CDATA[[好的]]]></tw-value>
<th-value><![CDATA[[NoProb]]]></th-value>
<fileName>NoProb.png</fileName>
<eggIndex>2034</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[MyBad]]]></key>
<cn-value><![CDATA[[打脸]]]></cn-value>
<qq-value><![CDATA[[打脸]]]></qq-value>
<en-value><![CDATA[[MyBad]]]></en-value>
<tw-value><![CDATA[[打臉]]]></tw-value>
<th-value><![CDATA[[MyBad]]]></th-value>
<fileName>Slap.png</fileName>
<eggIndex>2036</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[Wow]]]></key>
<cn-value><![CDATA[[哇]]]></cn-value>
<qq-value><![CDATA[[哇]]]></qq-value>
<en-value><![CDATA[[Wow]]]></en-value>
<tw-value><![CDATA[[哇]]]></tw-value>
<th-value><![CDATA[[Wow]]]></th-value>
<fileName>Wow!.png</fileName>
<eggIndex>2037</eggIndex>
</emoji>
<emoji>
<key><![CDATA[[KeepFighting]]]></key>
<cn-value><![CDATA[[加油加油]]]></cn-value>
<qq-value><![CDATA[[加油加油]]]></qq-value>
<en-value><![CDATA[[KeepFighting]]]></en-value>
<tw-value><![CDATA[[加油!]]]></tw-value>
<th-value><![CDATA[[KeepFighting]]]></th-value>
<fileName>KeepFighting.png</fileName>
<eggIndex>2038</eggIndex>
</emoji>
</newemoji>
Binary file not shown.

After

Width:  |  Height:  |  Size: 2.1 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.0 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.7 KiB

+67 -11
View File
@@ -5,14 +5,16 @@ import xml.etree.ElementTree as ET
import os
import json
tree = ET.parse(os.path.join(
os.path.dirname(os.path.abspath(__file__)), 'smiley.xml'))
root = tree.getroot()
smileys = {}
for child in root:
name = child.attrib['name']
if 'smiley_values' in name:
def parse_smiley_xml():
ret = {}
tree = ET.parse('smiley.xml')
root = tree.getroot()
for child in root:
name = child.attrib['name']
if 'smiley_values' not in name:
continue
if '_th' in name:
continue # ignore thailand language
lst = [c.text for c in child]
@@ -23,9 +25,63 @@ for child in root:
v = v.strip('"')
v = v.replace('&lt;', '<')
v = v.replace('&amp;', '&')
v = v.decode('utf-8')
smileys[v] = idx
ret[v] = os.path.join("smileys", f"{idx}.png")
return ret
with open('tencent-smiley.json', 'w') as f:
json.dump(smileys, f)
def parse_extra_smiley():
# some extra smiley from javascript on wx.qq.com
with open("tencent-smiley-extra.json") as f:
obj = json.load(f)
extra = {'[' + k + ']': os.path.join("smileys", f"{v}.png") for k, v in obj.items()}
return extra
def parse_new_emoji():
ret = {}
xmlfile = "newemoji/newemoji-config.xml"
tree = ET.parse(xmlfile)
root = tree.getroot()
for child in root:
assert child.tag == "emoji"
vals = {k.tag: k.text for k in child}
filename = os.path.join("newemoji", vals["fileName"])
for k, v in vals.items():
if "-value" in k:
ret[v] = filename
return ret
def parse_unicode_smiley():
# 1f35c -> "\ue340"
#self.unicode_smiley_code = gUnicodeCodeMap
# u'\U0001f35c' -> "e340" # for iphone
# u'\ue415' -> 'e415' # for android
def unichar(i):
try:
return chr(i)
except ValueError:
return struct.pack('i', i).decode('utf-32')
ret = {}
with open("unicode-smiley.json") as f:
d = json.load(f)
for k, v in d.items():
fname = os.path.join("smileys", hex(ord(v))[2:] + ".png")
ret[unichar(int(k, 16))] = fname
ret[v] = fname
return ret
if __name__ == "__main__":
# parse old smileys
smileys = {}
def add(dic, name):
smileys.update(dic)
print(f"Found {len(dic)} smileys from {name}. Total is {len(smileys)}")
add(parse_smiley_xml(), "smiley.xml")
add(parse_extra_smiley(), "tencent-smiley-extra.json")
add(parse_new_emoji(), "newemoji")
add(parse_unicode_smiley(), "unicode-smiley.json")
with open('tencent-smiley.json', 'w') as f:
json.dump(smileys, f)
File diff suppressed because one or more lines are too long