Rapido web service branche - Pull request #69

anki commented on 19 Dec 2023

Pull request for rapido web service source code.
Task remaining :

Comment the code
Readme
Test on VI so further change may be coming on some files

parmentf commented on 19 Dec 2023

~~I think you should name it using a more specialized name than rapido, which applies to all services in data-computer.~~
I was wrong.

data-computer/v1/charger-xml.cfg

	25	value = get(env('pid')).replace('uid:/', '')
	26
	27	[metrics]
	28	bucket = charger

data-computer/v1/rapido-ws.ini

	2	mimeType = application/json
	3	# OpenAPI Documentation - JSON format (dot notation)
	4	post.operationId = post-v1-rapido-ws
	5	post.description = WS-rapido

parmentf on 19 Dec 2023

I can't understand the operation done by the web service, only by reading its description (nor its summary).

data-computer/v1/rapido-ws.ini

	34	plugin = basics
	35	plugin = spawn
	36	# Step 1 (générique): Charger le fichier corpus
	37

parmentf on 19 Dec 2023

Step 1, Step 2, and Step 2.1 comments should be nearer to the instruction they comment.
Otherwise, we could believe that they comment the lines above them.

data-computer/v1/rapido-ws.py

	2
	3	import sys
	4	import json
	5	import sys

parmentf on 19 Dec 2023

Imported twice !

anki commented on 19 Dec 2023

The problem is that this web service will probably only be used by Persee as part of the rapido project. It will probably be easier for Persee people to use the web service if it evokes the project name rather than a vague description of what the web service does, such as "idref-alignment". Or at least include rapido in the name: "Rapido-idRef-alignment", but that's a long name.

parmentf commented on 19 Dec 2023

Sorry, I thought of TDM Factory, instead of Rapido, which is a project name.
I forgot this.
But you can remove the -ws part from the name: it is a web services repository ;)

data-computer/v1/rapido-ws.py

	72
	73	for line in sys.stdin:
	74	data = json.loads(line)
	75	print(data,file=sys.stderr

parmentf on 19 Dec 2023

Isn't there too much debug traces?

data-computer/v1/rapido-ws.py

	76	)
	77	print("RECEIVE",file=sys.stderr)

parmentf on 19 Dec 2023

At least, add a timestamp for a useful log.

data-computer/v1/rapido/preprocessing.py

	1	import re
	2	import os

parmentf commented on 19 Dec 2023

I did not read the whole code, but it seems reasonable (besides the is not None convention (instead of != None).

anki referenced the pull request on 22 Dec 2023

Rapido web service branche

anki merged commit 24926dc into master from rapido-data-computer on 22 Dec 2023

anki closed this pull request on 22 Dec 2023

anki deleted the rapido-data-computer branch on 22 Dec 2023

Rapido web service branche #69