chore: optimize performance (replace he with entities)

[oppilate]

Before:

➜ ab -k -c 20 -n 250 http://127.0.0.1:1200/initium/feature/zh-hant           
This is ApacheBench, Version 2.3 <$Revision: 1879490 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 127.0.0.1 (be patient)
Completed 100 requests
Completed 200 requests
Finished 250 requests


Server Software:        
Server Hostname:        127.0.0.1
Server Port:            1200

Document Path:          /initium/feature/zh-hant
Document Length:        224091 bytes

Concurrency Level:      20
Time taken for tests:   42.454 seconds
Complete requests:      250
Failed requests:        0
Keep-Alive requests:    250
Total transferred:      56123724 bytes
HTML transferred:       56022750 bytes
Requests per second:    5.89 [#/sec] (mean)
Time per request:       3396.299 [ms] (mean)
Time per request:       169.815 [ms] (mean, across all concurrent requests)
Transfer rate:          1291.01 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.2      0       1
Processing:   131 2773 580.9   2786    6833
Waiting:      131 2772 580.8   2786    6832
Total:        131 2773 580.7   2786    6833

Percentage of the requests served within a certain time (ms)
  50%   2786
  66%   2927
  75%   3030
  80%   3120
  90%   3341
  95%   3400
  98%   3540
  99%   3547
 100%   6833 (longest request)

After:

➜ ab -k -c 20 -n 250 http://127.0.0.1:1200/initium/feature/zh-hant
This is ApacheBench, Version 2.3 <$Revision: 1879490 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking 127.0.0.1 (be patient)
Completed 100 requests
Completed 200 requests
Finished 250 requests


Server Software:        
Server Hostname:        127.0.0.1
Server Port:            1200

Document Path:          /initium/feature/zh-hant
Document Length:        224164 bytes

Concurrency Level:      20
Time taken for tests:   14.996 seconds
Complete requests:      250
Failed requests:        0
Keep-Alive requests:    250
Total transferred:      56141974 bytes
HTML transferred:       56041000 bytes
Requests per second:    16.67 [#/sec] (mean)
Time per request:       1199.655 [ms] (mean)
Time per request:       59.983 [ms] (mean, across all concurrent requests)
Transfer rate:          3656.13 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    0   0.4      0       3
Processing:    61  704 374.9    650    6173
Waiting:       60  704 374.8    650    6170
Total:         61  704 374.9    650    6173

Percentage of the requests served within a certain time (ms)
  50%    650
  66%    686
  75%    710
  80%    734
  90%    877
  95%   1029
  98%   1101
  99%   1102
 100%   6173 (longest request)

[vercel]

This pull request is being automatically deployed with Vercel (learn more).
To see the status of your deployment, click below or on the icon next to each commit.

🔍 Inspect: https://vercel.com/diy/rsshub/kjxy7ezkd
✅ Preview: https://rsshub-git-fork-oppilate-speedup.diy.vercel.app

[DIYgod]

entities should be moved from devDependencies to dependencies

RSSHub/package.json

Line 44 in 18adc60

"entities": "2.1.0",

[oppilate]

Wired. So yarn add entities won't add it to dependencies if it's already in devDependencies. Fixed now.

[oppilate]

And parse5 used by cheerio accounts for another 1/3 processing time. Cheerio is capable of handling htmlparser2's parsed dom tree, which is reportedly 5 times faster than parse5. But that requires one more line in each cheerio parsing call:

// Usage as of htmlparser2 version 3:
const htmlparser2 = require('htmlparser2');
const dom = htmlparser2.parseDOM(document, options);

const $ = cheerio.load(dom);

I won't do it right now. Just to take a note here.

[HenryQW]

htmlparser2 is the default now.

cheeriojs/cheerio#985

[oppilate]

htmlparser2 is the default now.

Well, actually it's htmlparser2 for XML and parse5 for HTML: cheeriojs/cheerio#866 (comment). In RSSHub only the HTML part is used (if I didn't miss anything), which matches what I see in the profiling result.

Anyway, that's just a 3/2 speed up, not so dramatic like 3 times speedup in this PR.