[Nozomi] Database feature doesn't work properly for Nozomi

[oxi7589]

Looks like Nozomi submissions are not uniquely represented in the database, making this feature largely useless for this extractor. The problem is reproducible on current stable version, 1.17.3

Config:

    "nozomi": {
        "archive" : "./db/nozomi.sqlite3",
        "filename": "{postid}-{num} {dataid}.{extension}"
    },

Sample output (assuming ./db/nozomi.sqlite3 is blank or absent):

PS U:\gallery-dl> .\gallery-dl "https://nozomi.la/search.html?q=pixiv_id_4169977"
* .\gallery-dl\nozomi\pixiv_id_4169977\27699823-1 8e3c15aa266fc581f72d24af68ffd26770c948e4ef8effa8610c229cf540f15b.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27699823-2 75275890de3e21cbe23bc5021a5085c2523d34a9eb3b85586214e5405441d642.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27699823-3 b50aaae9ad690f90b1d9f0a0c91333a974d2dd012f3b6c5ed155988ef39d7fe2.png
* .\gallery-dl\nozomi\pixiv_id_4169977\27652373-1 82af69c9438597b190ee0b43e6d74a2a8e313ebeb997d66ca40944083cf08367.mp4
* .\gallery-dl\nozomi\pixiv_id_4169977\27616177-1 0b1c1ae162ba1085808c6043fe0992aed7407f0723421f002157488c0982d44c.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27616177-2 cd911f8e297cfe85f70974bd575f443bdfa475343969cd36b59920505fac1d5e.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27616177-3 763fd9c528616be3428e092c90c0c4e197833019080ab68dc9e5b0d81adebb52.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27616177-4 d416ab9c958f8b267087c2d981ec39b299ef558c3ff890fee0b0baea9e7217bd.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27616177-5 0689d6ec7f93cbb848db2ed19b6fa57d8015fad6fb3bb85c6c2f1095e3a8b12a.png
* .\gallery-dl\nozomi\pixiv_id_4169977\27549156-1 459245d1401e62476bd4841cb638a30b4708af7982cf9a3a791a5a1123a902d4.png
# .\gallery-dl\nozomi\pixiv_id_4169977\27549156-2 596020ee52ec01fb13651a58b1b1979c66c58fe0cb9d163af9a9a58f2068d133.png
...

Notice how only the 1st file in a sequence is processed properly.

[Hrxn]

Might be caused by the archive format setting for Nozomi.

Note: Filename settings are entirely irrelevant here.

I think in your example, NozomiSearchExtractor gets used, which has a default archive format of "t_{search_tags}_{postid}"

Maybe that's the problem, because if I recall it correctly, Nozomi is a site that breaks with typical booru conventions, so that one post can have multiple entries/images?

Could you try it with a config like this, for example?

"nozomi": {
        "search": {
              "archive" : "./db/nozomi.sqlite3",
              "filename": "{postid}-{num} {dataid}.{extension}",
              "archive-format": "t_{search_tags}_{postid}_{dataid}"
         }
    },

[oxi7589]

"archive-format": "t_{search_tags}_{postid}_{dataid}"

Yes, this solves the problem. Maybe this setting should be the default for this site, since the current default is not exactly functional.