List all linked-to nonexistent files

[Gewerd-Strauss]

Hello,

I'd like to get an overview over missing links in my vault, and found the following DVJS-query for doing so:

//how many links a non existing file should have at minimum
const count = 2;

//specify the full path here.
const ignoredExisting = ["your/ignored/notes/here.md"];

//keep these in lower case.
const ignoredNonExisiting = ["your non exisiting notes", "here is note that does not exist"];

let d = {};
function process(k, v) {
  Object.keys(v).forEach(function (x) {
    if(!ignoredNonExisiting.includes(x.toLowerCase())) {
        x = dv.fileLink(x);
        if (d[x]==undefined) { d[x] = []; }
        if(!ignoredExisting.includes(k)) {
            d[x].push(dv.fileLink(k));
        }
    }
  });
}

Object.entries(dv.app.metadataCache.unresolvedLinks)
    .filter(([k,v]) => Object.keys(v).length)
    .forEach(([k,v]) => process(k, v));
    
dv.table(["Non existing notes", "Linked from"],
         Object.entries(d)
         .filter(([k, v]) => v.length >= count)
	     .sort((a, b) => b[1].length - a[1].length)
         .map(([k,v]) => [k, v.join(" • ")]));

In theory, this would be as simple as setting count=1. But... that results in 573 entries. Which is.... way too many, and a lot of them are meaningless files I don't care about in the scope of this.

Because it is too unprecise, there are certain subfolders and/or filenames I want to exclude. And, as far as I could test, adding any filename to ignoredNonExisting does not actually exclude them.

Query taken from DataviewJS Snippet Showcase on the forums.

So what I need help with is equipping this snippet with

1. a way to filter out files based on their paths, partial paths (e.g. only folder names)
2. a way to filter them out by name

And preferably an array because it makes extending it so much easier than hardcoding, for obvious reasons.

Would someone know how to adjust this code?

Thank you,
Sincerely,
~Gw

[AB1908]

What do you mean by "missing links"? Do you mean links to non-existent files?

[Gewerd-Strauss]

I am okay with that for this sunday, or either day next weekend - not today. I am busy on other fronts right now and can't pause that today.

[Gewerd-Strauss]

I don't know what timezone you're in, so what's your timeframe tomorrow in CEST/GMT+2?

[AB1908]

Anything after 9 AM GMT should be fine for me.

[Gewerd-Strauss]

Plans never survive first contact with reality. I wanted to be done by 11 AM GMT, guess that is gone now. How shall I contact you on the DC? I am not terribly active on it, do I just ping you? If so, where?

[AB1908]

Do try the dangling links plugin in case that helps.

[mariomui]

/*
unresolvedLinks: Record<
    SourceFilename extends string, 
    Record<
        brokenLinkName extends string, 
        brokenLinkCount extends number
        >
    > as BrokenLinkNameToCount

const exampleUnresolvedLinksValue: BrokenLinkNameToCount = {
 "i_dont_exist.md" : 1
}
A) currently: dv.app.metadataCache.unresolvedLinks  = {
    "insert_file_path": NonEmptyObject, 
    "other_stuff": EmptyObject
} 
B) i-want-a-dataview-proper-format:
    unresolvedLinks : [ [NonEmptyObject.key, nonEmptyObject.values], ... ]   
    // other_stuff and its EmptyObject is removed. 
*/

// --- DATA LAYER BOUNDARY
// type unresolvedTuples: UnResolvedTuple[]
// type UnresolvedTuple = [sourceFileName, BrokenLinkNameToCount] where BrokenLinkNameToCount is nonEmpty
// purpose: exclude empty objects from our unresolved tuples list.
const unresolvedLinkTuples = Object.entries(
  dv.app?.metadataCache?.unresolvedLinks || {}
);

// --- BUSINESS LOGIC BOUNDARY
const thresholdedUnresolvedTuples = unresolvedLinkTuples.filter(
  ([sourceFileName, brokenLinkNameToCountMapping]) => {
    return Object.keys(brokenLinkNameToCountMapping).length;
  }
);
const transformedTuples = thresholdedUnresolvedTuples;

// RENDERING LOGIC BOUNDARY
const sourceFileNameAndLinkNamesTuples = transformedTuples.map(
  ([sourceFileName, linkNameToCountContext]) => {
    const linkNames = Object.entries(linkNameToCountContext).map(
      keepOnlyKeyFromTupleMapPredicate
    );
    return [sourceFileName, linkNames];
  }
);
const sourceFileLinkAndLinkNamesTuples = sourceFileNameAndLinkNamesTuples.map(
  ([sourceFileName, linkNames]) => {
    const sourceFileLink = dv.fileLink(sourceFileName);
    return [sourceFileLink, linkNames];
  }
);

const headers = ["source-link", "non-existent-link-name"];
dv.table(headers, dv.array(sourceFileLinkAndLinkNamesTuples));

function keepOnlyKeyFromTupleMapPredicate([key]) {
  return key;
}

I think its easier to read if the logic is split up, and use better variable names. i can't read k, v stuff i wrote after 2 months later.
For the count threshold, thats the filter for the unresolvedTuples.

Then, for the ignorePaths then you would iterate over each keyvalue tuple of your unresolvedTuples, grab the valueTuple which contains a RecordType, grab the key of that recordType which contains the full path of the unresolved link, and filter out any fullpath that is included in your filepaths. ignorePaths.some(ignorePath => brokeLinkpath.startsWith(ignorePath)

Because the starting object is unresolved links from the cache, its not gonna be huge. There's no need to push all the logic together.
its all transforming logic in the end. The only contractual obligation is the sourceFilename being NON-transformed, allowing you to focus your code on the value Tuple of the equation.
i forgot to mention that i didn't implement your requirements. I just set up a guilded rose type of refactor so you could just insert your own transforming logic in a reasonable way.