# Developing custom scanner, when does Scan() recursion happen?

**URL:** <https://forums.plex.tv/t/developing-custom-scanner-when-does-scan-recursion-happen/728067>\
**Category:** Dev/API Corner\
**Tags:** scanner-agent-dev\
**Created:** [July 11, 2021, 12:32pm UTC](https://forums.plex.tv/t/developing-custom-scanner-when-does-scan-recursion-happen/728067 "2021-07-11T12:32:38Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![waldosax](https://sea1.discourse-cdn.com/plex/user_avatar/forums.plex.tv/waldosax/32/246245_2.png) [@waldosax](https://forums.plex.tv/u/waldosax)\
**Post date:** [July 11, 2021, 12:32pm UTC](https://forums.plex.tv/t/developing-custom-scanner-when-does-scan-recursion-happen/728067/1 "2021-07-11T12:32:39Z")

</div>

# Developing custom scanner, when does Scan() recursion happen?

I’m developing a custom scanner/metadata agent as kind of an exercise to teach myself Python.

I’ve found a handful of resources floating around the web, but no comprehensive documentation about scanning, so I’m trying to reverse-engineer my scanner based on other scanners I’ve found.

I suppose my biggest question for the moment is, when/where/how many times is the Scan method called?

I’ve found, in a number of places, the following test code:

```auto
if __name__ == ' __main__': #command line
  path = sys.argv[1]
  files = [os.path.join(path, file) for file in os.listdir(path)]
  media = []
  Scan(path[1:], files, media, [])
  print("Files detected: ", media)

```

If I point this to my own library structure, I could potentially get varying results. I’m just trying to understand.

So given this folder structure:

```auto
/
	$PLEX_HOME/Library/Application Support/Plex Media Server/	<-- Plex main folder
	(%USERPROFILE%\Local Settings\Application Data\Plex Media Server\) <-- Plex main folder on Windows
	(/var/lib/plexmediaserver/Library/Application Support/Plex Media Server/) <-- Plex main folder on Debian
		Plug-ins/
			PlexSportsAgent.bundle/
				Contents/
					Code/
						__init__.py <-- Metadata Agent script
					Resources/
					DefaultPrefs.json
					info.plist
		Scanners/
			Series/
				PlexSportsScanner/	<-- Scanner script (not sure if I can put in a folder structure like this, but I'd like to keep it organized)
					Data/
						Leagues/	<-- Should this be moved somewhere else?
							MLB/
								Teams.json	<-- Programmatically cached teams info
							NBA/
								Teams.json	<-- Programmatically cached teams info
							NFL/
								Teams.json	<-- Programmatically cached teams info
							NHL/
								Teams.json	<-- Programmatically cached teams info
						__init__.py
						SportsDataIO.py
						Teams.py
						TheSportsDB.py
					__init__.py <-- Defines the Scan method
	mnt/Media/
		Video/
			Sports/ <-- Sports library
				MLB/
					2021/
				NBA/
				NFL/
					2004-2005/
						NFL.Super Bowl.XXXIX.Patriots.vs.Eagles.720p.HD.TYT.mp4
						NFL.Super Bowl.XXXIX.Patriots.vs.Eagles.720p.HD.TYT.ts
					2017-2018/
						Super.Bowl.LII.2018.02.04.Eagles.vs.Patriots.1080p.HDTV.x264.Merrill-Hybrid-5.1-PHillySPECIAL.mkv
				NHL/
				UFC/
				.plexignore
				Phillies vs. Red Sox Game Highlights (7_10_21) _ MLB Highlights.mp4
				yt1s.com - Phillies vs Cubs Game Highlights 70821 MLB Highlights.mp4
				yt1s.com - Phillies vs Red Sox Game Highlights 7921 MLB Highlights.mp4

```

… the aforementioned script yields me the following results, presuming that `sys.argv[1]` was the path to the root of my Sports Library

```auto
Scan(path, files, media, subdirs, root=None)

path = "mnt/Media/Sports"
files = [
	"/mnt/Media/Sports/MLB",
	"/mnt/Media/Sports/NBA",
	"/mnt/Media/Sports/NFL",
	"/mnt/Media/Sports/NHL",
	"/mnt/Media/Sports/UFC",
	"/mnt/Media/Sports/.plexignore",
	"/mnt/Media/Sports/Phillies vs. Red Sox Game Highlights (7_10_21) _ MLB Highlights.mp4",
	"/mnt/Media/Sports/yt1s.com - Phillies vs Cubs Game Highlights 70821 MLB Highlights.mp4",
	"/mnt/Media/Sports/yt1s.com - Phillies vs Red Sox Game Highlights 7921 MLB Highlights.mp4"
]
subdirs = []

```

My assumptions here are that:

- `sys.argv[1]` is **fully-qualified** (just for argument’s sake)
- `files` only has the **surface depth** of `path`. If so, when does the recursion take place? The way this is currently set up, this will script will yield 5 directories and 4 files (Phillies games and a `.plexignore` file)  
– If I pass this through to my own `Scan()`, which in turn, gets filtered by `VideoFiles.Scan()`, ultimately I am left with 3 video files to chomp on.  
– Am I responsible for recursing the remaining folder structures myself? Or is `Scan()` called multiple times from the outside for the next level of depth?  
– If I am responsible for recursing myself, should the `subdirs` parmeter be populated? Presumably with the 5 subfolders? Or should I take the test script at face value and discover the subfolders from the `files` parameter? Should the `files` parameter be filtered from the test script to only include files, not files AND folders (`os.listdir(path) if os.path.isfile(file)` - _or something like that_)

As it currently stands, `Scan()` will only yield me 3 Phillies games, and whack everything else out of existence. This is why I’m confused.

Can someone, perhaps a member of the Plex team, assist me in understanding?

---

<div class="post-metadata">

**Author:** ![saltydk](https://avatars.discourse-cdn.com/v4/letter/s/a698b9/32.png) [@saltydk](https://forums.plex.tv/u/saltydk)\
**Post date:** [July 25, 2021, 3:35pm UTC](https://forums.plex.tv/t/developing-custom-scanner-when-does-scan-recursion-happen/728067/2 "2021-07-25T15:35:21Z")

</div>

Try looking at [Absolute-Series-Scanner/Absolute Series Scanner.py at 3188799d714c2334c5990a8e2f7508cb0c3fbc38 · ZeroQI/Absolute-Series-Scanner · GitHub](https://github.com/ZeroQI/Absolute-Series-Scanner/blob/3188799d714c2334c5990a8e2f7508cb0c3fbc38/Scanners/Series/Absolute%20Series%20Scanner.py#L490)

My guess is that you need to call it on every root directory.

---

<div class="post-metadata">

**Author:** ![waldosax](https://sea1.discourse-cdn.com/plex/user_avatar/forums.plex.tv/waldosax/32/246245_2.png) [@waldosax](https://forums.plex.tv/u/waldosax)\
**Post date:** [July 27, 2021, 4:37pm UTC](https://forums.plex.tv/t/developing-custom-scanner-when-does-scan-recursion-happen/728067/3 "2021-07-27T16:37:29Z")

</div>

After tinkering with it a bit, I have come to this same conclusion as well.