Search Query builder 1/2 - Example

[Gaisberg]

From Discord (by monika_10356):

Search Category:

Is there a possibility to configure the category you're giving to the indexer in the search query based on type and device? 
Then i can be pretty sure it will give me the best results.
Im from Germany and my indexer has an extra movie-de category 2100 for example. 
Movie-es (spanish) would be 2200 (see screen category.png)
Same goes for TV. TV-DE is 5100.

Search Query:
If you could add the option to add things to the search query, then i could pre select and reduce the results.
You can search for groups or search for resolutions or which kind of rip.

Example search:
searching over all categories

• search: "imdbid=27543632" -> gives me 72 results

searching with the right category 5100 (Movies-DE)

• search: "the+housemaid" -> gives me 61 results (there is also a housmaid from 2021)
• search: "the+housemaid+2025" -> gives me 44 results
• search: "the+housemaid+2025+P73" -> gives me two results
• search: "the+housemaid+2025+(P73+OR+LRO)" -> gives me 10 results (| would also work as OR)
• search: "the+housemaid+2025+(P73|LRO)+(1080p)" -> gives me 4 results
  I could also pre select only 1080p and 2160p and so on

At least on the webpage you can combine id and query (API maybe too?)

• search: "imdbid=27543632" -> gives me 44 results
• search: "imdbid=27543632&q=(P73)" -> gives me two results
• search: "imdbid=27543632&q=(P73+OR+LRO)" -> gives me 10 results
  and so on

[Gaisberg]

gaisberg (Discord):

Would something like this suffice?

[Gaisberg]

monika_10356 (Discord):

yes
i think this is great and the fastest way for an implementation.
Search builder like mentioned would be better for others to understand.
But a single line of input is enough

but wouldnt it be better to do this per device instead of indexer?

example:
letzt say one part of my family only needs german content

and the other part would also like english content or another part of the family only wants spanish content.

i could also pre select quality based on the device. The TV of the parents would not need anything higher than hd material.

my home cinema device would like 4k or FHD

mobile ...
and so on

[Gaisberg]

gaisberg (Discord):

Yeah definately, give :main a go and I'll look into implementing this per device instead of per indexer

[Gaisberg]

monika_10356 (Discord):

thanks
will test tomorrow 🙂

[Gaisberg]

monika_10356 (Discord):

i would say there is something wrong 😅

i tried with "Mercy 2026"

msg="searchAndValidate" imdb= tvdb= season= ep= maxStreams=6
msg="AvailNZB GetReleases" imdb_id=tt31050594 tvdb_id= season=0 episode=0 indexer_filter=1
msg="AvailNZB GetReleases" count=17 available=15 imdb_id=tt31050594 tvdb_id=
msg="AvailNZB releases" total=17
msg="AvailNZB phase" releases=15 after_triage=1
msg="AvailNZB phase done" streams=1
msg="Newznab search request" indexer=SceneNZBs url=https://scenenzbs.com/api?apikey=redacted&cat=2100&limit=1000&o=xml&offset=0&q=%28FuN%7CP73%7CVECTOR%29&t=movie&tmdbid=1236153 limit=1000
msg="Newznab search request" indexer=SceneNZBs url=https://scenenzbs.com/api?apikey=redacted&cat=2100&limit=1000&o=xml&offset=0&q=Mercy+%28FuN%7CP73%7CVECTOR%29&t=movie limit=1000
msg="Indexer dual search" id=500 text=227
msg="Indexer candidates after triage" count=407```

what is the last limit =1000 after each request? is it just showing how much was limited in the search request or is it also a part of the search request?

i try with the api in the Browser and give you detailed results in a moment.

[Gaisberg]

monika_10356 (Discord):

Implemented right now:

https://scenenzbs.com/api?apikey=redacted&cat=2100&limit=1000&o=xml&offset=0&q=Mercy+%28FuN%7CP73%7CVECTOR%29&t=movie
<newznab:response offset="0" total="367"/> wrong


https://scenenzbs.com/api?apikey=redacted&cat=2100&limit=1000&o=xml&offset=0&q=%28FuN%7CP73%7CVECTOR%29&t=movie&tmdbid=1236153

<newznab:response offset="0" total="612"/> wrong

[Gaisberg]

monika_10356 (Discord):

Test api calls:

https://scenenzbs.com/api?apikey=redacted&t=movie&tmdbid=1236153

<newznab:response offset="0" total="84"/> correct


https://scenenzbs.com/api?apikey=redacted&t=search&tmdbid=1236153

<newznab:response offset="0" total="84"/> correct

https://scenenzbs.com/api?apikey=redacted&t=movie&cat=2000,2100&tmdbid=1236153

<newznab:response offset="0" total="84"/> correct


https://scenenzbs.com/api?apikey=redacted&t=search&cat=2000,2100&tmdbid=1236153

<newznab:response offset="0" total="84"/> correct

https://scenenzbs.com/api?apikey=redacted&t=movie&cat=2000&tmdbid=1236153

<newznab:response offset="0" total="45"/> correct


https://scenenzbs.com/api?apikey=redacted&t=search&cat=2000&tmdbid=1236153

<newznab:response offset="0" total="45"/> correct

https://scenenzbs.com/api?apikey=redacted&t=movie&cat=2100&tmdbid=1236153

<newznab:response offset="0" total="39"/> correct

https://scenenzbs.com/api?apikey=redacted&t=movie&cat=2100&limit=1000&o=xml&offset=0&tmdbid=1236153
<newznab:response offset="0" total="39"/> correct

https://scenenzbs.com/api?apikey=redacted&t=search&cat=2100&q=Mercy%20%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="12"/> also includes different movies with mercy in the name because there is no year in the name. Filters are correctly used.

https://scenenzbs.com/api?apikey=redacted&t=search&cat=2100&q=Mercy%202026%20%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="1"/> correct

[Gaisberg]

monika_10356 (Discord):

Result of my test:

• You have to use t=search instead of t=movie
• Best results are with movie title + year + filter and not with id + filter. (There was a wrong movie always showing up, because it had the tmdbid from Mercy. It’s already corrected from the indexer. But this shows that filename is the best search indicator)

[Gaisberg]

monika_10356 (Discord):

Additional requests:

• let us set the Limit value.
• use name + year for search.
• An option to use the german name from imdb or tmdb would also be great. But here needs to be some normalisation of some german additional letters. (e.g. ü = ue, ö =oe, and so on. English title=Lion King, German title= König der Löwen, File Name= Koenig der Loewen) maybe umlautadaptarr can help with these. I think other languages have these problems too

Biggest change idea:
can you give us some kind of table in which we can define how we want to search?

Table data for search priority:
id or original name or translated normalized name, year, category, serach filter, availnzb, sorting, result filter, result Limit

lets say i have three entries:

1. translated normalized name(german), year, category, search filter, result limit(10)
2. translated normalized name(german), year, category, availnzb, sorting, result filter, result Limit(50)
3. id, year, category, search filter, availnzb, sorting, result filter, result limit(100)

in the case of Mercy 2026 it would start with the first entry and also stops right there. Because it will give one return.
If the first search did not find a result. It will try the second (if there is a second of corse).

with this setup it should be the absolut best results

[Gaisberg]

monika_10356 (Discord):

these need to be also per device

maybe i want to first find the best german results and the second search would be the englisch results

[Gaisberg]

monika_10356 (Discord):

I forgot to test with TV Show.

i tried with Fallout 2024 Season 1 Episode 1

Implemented right now:
/api?apikey=redacted&cat=5100&ep=1&limit=1000&o=xml&offset=0&q=Fallout+S1E1+%28FuN%7CP73%7CVECTOR%29&season=1&t=tvsearch
<newznab:response offset="0" total="0"/> wrong. You have to search with S01E01

/api?apikey=redacted&cat=5100&ep=1&limit=1000&o=xml&offset=0&q=%28FuN%7CP73%7CVECTOR%29&season=1&t=tvsearch&tvdbid=416744
<newznab:response offset="0" total="4"/> correct

Test API calls:
/api?apikey=redacted&t=tvsearch&cat=5100&season=1&ep=1&limit=1000&o=xml&offset=0&q=Fallout+S01E01+%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="4"/>

/api?apikey=redacted&t=tvsearch&cat=5100&limit=1000&o=xml&offset=0&q=Fallout+S01E01+%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="4"/>

/api?apikey=redacted&t=search&cat=5100&limit=1000&o=xml&offset=0&q=Fallout+S01E01+%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="4"/>

Result of my test:

• t=tvsearch is correct. no need to use search. But t=search will give the same results
• i think &season=1&ep=1 should be configurable too. If we want to use these or not. these are not always filled
• year is sometimes used and sometimes not in the filename
• Episode numbering for TVshows is in the Format SxxExx
• Anime numbering is sometimes like mentioned above and sometimes absolute. (e.g. you can find files with One Piece 1999 E0001 and One Piece 1999 S01E01)

The mentioned Search priority table can help with these too

[Gaisberg]

gaisberg (Discord):

Could you draft a search priority table that you would like to see?

[Gaisberg]

monika_10356 (Discord):

sure will do if im back home

[Gaisberg]

gaisberg (Discord):

Also some changes in main image for you to test out

[Gaisberg]

monika_10356 (Discord):

i contacted the support from scene and they fixed some bugs in the API and updated their Documentation accordingly. 😅

tldr:

/api?apikey=redacted&t=movie&cat=2100&tmdbid=1236153&q=%28FuN%7CP73%7CVECTOR%29
<newznab:response offset="0" total="1"/> correct
this bug is fixed and works now according to the newznab standard

Use the following two templates for movie search:

• t=movie&cat=2100&tmdbid=1236153&q=%28FuN%7CP73%7CVECTOR%29
• t=search&cat=2100&q=Mercy+2026+%28FuN%7CP73%7CVECTOR%29

and the following two templates for tv show search:

• t=tvsearch&cat=5100&tvdbid=416744&season=1&ep=1&q=%28FuN%7CP73%7CVECTOR%29
• t=search&cat=5100&q=Fallout+S01E01+%28FuN%7CP73%7CVECTOR%29