File finding capabilities fluctuates

[olivierroymd]

Hello,

I rely on Alfred a lot for my work. I handle a lot of PDFs and it's the primary reason I use Alfred (for quickly opening PDFs).

 

A few weeks ago, I started noticing that my PDFs could not be found even though I changed absolutely nothing in my workflow except that my PDF library changed location on my computer (was previously in my Google Drive and now located in /user/Zotero/storage/random_folderID/*.pdf). I can always find the files using Spotlight. I rebuilt my Spotlight index several times but did not do anything. I have reinstalled Alfred and reseted the preferences, no effect. I have cleared Alfred cache several times, no effect. I also cleared the Alfred knowledge several times, no effect. The folders that contains PDF is in the Search scope. I asked the Zotero guys and they insisted this had nothing to do with them. I have searched a lot and found out that this may have to do with the recency of use of the file and might be related to some extended attributes of the files. However, the metadata checker built-in with Alfred shows no problem:

 

For example:

 

Starting Diagnostics...

File: 'PotvinO_2017_NeuroImage.pdf'
Path: '/Users/olivierroy/Zotero/storage/6P9LWBDL'

-----------------------------------------------------------

Check file cache database...

✅ File cache integrity is ok

-----------------------------------------------------------

Check if file is readable...

✅ Alfred has permissions to read this file.

-----------------------------------------------------------

Check if volume '/' is indexed by macOS...

✅ Indexing is enabled on this drive

-----------------------------------------------------------

Check direct file metadata...

✅ Direct metadata available

Display Name: PotvinO_2017_NeuroImage.pdf
 Other Names: 
Content Type: com.adobe.pdf
   Last Used: 

-----------------------------------------------------------

Check mdls file metadata...

✅ Metadata contains required items

_kMDItemDisplayNameWithExtensions      = "PotvinO_2017_NeuroImage.pdf"
kMDItemAuthors                         = (
    "Olivier Potvin"
)
kMDItemContentCreationDate             = 2021-05-02 01:21:02 +0000
kMDItemContentCreationDate_Ranking     = 2021-05-02 00:00:00 +0000
kMDItemContentModificationDate         = 2021-07-09 09:37:00 +0000
kMDItemContentModificationDate_Ranking = 2021-07-09 00:00:00 +0000
kMDItemContentType                     = "com.adobe.pdf"
kMDItemContentTypeTree                 = (
    "com.adobe.pdf",
    "public.data",
    "public.item",
    "public.composite-content",
    "public.content"
)
kMDItemCreator                         = "DEVONthink 3"
kMDItemDateAdded                       = 2021-05-02 01:21:02 +0000
kMDItemDateAdded_Ranking               = 2021-05-02 00:00:00 +0000
kMDItemDescription                     = "NeuroImage, 156 (2017) 43-64. doi:10.1016/j.neuroimage.2017.04.035"
kMDItemDisplayName                     = "PotvinO_2017_NeuroImage.pdf"
kMDItemDocumentIdentifier              = 0
kMDItemEncodingApplications            = (
    "macOS Version 10.15.5 (Build 19F101) Quartz PDFContext"
)
kMDItemFinderComment                   = "zotero://select/library/items/GUEHRJEX"
kMDItemFSContentChangeDate             = 2021-07-09 09:37:00 +0000
kMDItemFSCreationDate                  = 2021-05-02 01:21:02 +0000
kMDItemFSCreatorCode                   = ""
kMDItemFSFinderFlags                   = 0
kMDItemFSHasCustomIcon                 = (null)
kMDItemFSInvisible                     = 0
kMDItemFSIsExtensionHidden             = 0
kMDItemFSIsStationery                  = (null)
kMDItemFSLabel                         = 0
kMDItemFSName                          = "PotvinO_2017_NeuroImage.pdf"
kMDItemFSNodeCount                     = (null)
kMDItemFSOwnerGroupID                  = 20
kMDItemFSOwnerUserID                   = 501
kMDItemFSSize                          = 1702019
kMDItemFSTypeCode                      = ""
kMDItemInterestingDate_Ranking         = 2021-07-13 00:00:00 +0000
kMDItemKeywords                        = (
    "Magnetic resonance imaging; Atrophy; Morphometry; Normality; Aging; Sex; Segmentation"
)
kMDItemKind                            = "Document PDF"
kMDItemLogicalSize                     = 1702019
kMDItemNumberOfPages                   = 22
kMDItemPageHeight                      = 793.701
kMDItemPageWidth                       = 595.2760009765625
kMDItemPhysicalSize                    = 1703936
kMDItemSecurityMethod                  = "None"
kMDItemTitle                           = "Freesurfer cortical normative data for adults using Desikan-Killiany-Tourville and ex vivo protocols"
kMDItemVersion                         = "1.3"

-----------------------------------------------------------

Check file is in search scope...

✅ File exists within Alfred's default search scope

-----------------------------------------------------------

✅ Troubleshooting passed
 

From now on, every now and then, I can't find any PDF. What I do to circumvent this issue is I run a simple bash script that simply apply the touch command to every PDF in my storage file.

for file in find /Users/olivierroy/Zotero/storage/*/*.pdf
do 
	touch -t 202107090537 $file
done

This resolves the issue...for a while. Then, usually after waking from sleep, the issue comes back.

 

My Alfred version is: Alfred 4.3.4 [1229]

My macOs version is: 11.4

 

I don't want to stop using Alfred and go back to Spotlight but I feel this will be the only solution going forward if I can't find a solution.

 

Thanks

[Andrew]

I wrote a kb article which is relevant to what you describe, where indexing is technically working, but the results fall outside of what is treated as relevant:

 

https://www.alfredapp.com/help/kb/understanding-result-ordering/

 

By running the touch in command line as you are, it will be updating the content changed date which will make the underlying macOS index highly prioritise this file for Alfred's search query (point 3 on that kb page).

 

What I would suggest in this case is to create a file filter workflow object which narrows down the search scope to within the folder you'd like to find, limit to pdfs and set the search limit up to 200 files. This way, any other files which the underlying macOS index is prioritising will get filtered out at source, and you'll be able to find the files you need.

[olivierroymd]

On 7/24/2021 at 5:03 AM, Andrew said:

I wrote a kb article which is relevant to what you describe, where indexing is technically working, but the results fall outside of what is treated as relevant:

 

https://www.alfredapp.com/help/kb/understanding-result-ordering/

 

By running the touch in command line as you are, it will be updating the content changed date which will make the underlying macOS index highly prioritise this file for Alfred's search query (point 3 on that kb page).

 

What I would suggest in this case is to create a file filter workflow object which narrows down the search scope to within the folder you'd like to find, limit to pdfs and set the search limit up to 200 files. This way, any other files which the underlying macOS index is prioritising will get filtered out at source, and you'll be able to find the files you need.

This unfortunately did not work.

It is very strange. I have to do the touch command every couple of hours.

Someone else got an idea?