What's the best way to search for specific terms in files?

Question

I have a situation where the client will be uploading a ton of pdf files, and then on the front-end we need a tool to search for key-terms within these files. Sometimes that key-term is in the file name (could possibly break apart with php to get the terms?) and sometimes the key-term is just in the file itself.

php to break apart the terms:

                <?php 
                    // break apart file name, grab first word
                    $str = '{title}'; 
                    $arr1 = preg_split("/[_,\- ]+/", $str); 
                ?>
                <?php echo $arr1[0]."\n"; ?>

Which takes a file name like: FFOBBULUSE000002_Harlequin GDP 3.5g TruFlower Cup.pdf and outputs the first term like: FFOBBULUSE000002

I've looked into Channel Files, and of course native EE search and File Entries, but none are really giving me what I need. Native search doesn't seem to be able to search native files, nor Channel Files?

Plus, there are thousands of these files, so I don't really want to load them all into a template and then filter. Hoping for a less resource-heavy way of doing this.

Example:

Uploaded file: FFOBULN2F000004_9lb 600mg Vape Cartridge.pdf

Someone may search for the term: FFOBULN2F000004 - that term needs to be able to pull up this file in the results, and also shouldn't be case sensitive.

Any ideas?

UPDATE:

I'm getting really close via using Stash. Here's what I have so far:

{exp:stash:set_list parse_tags="yes" name="my_entries"}

    {exp:file:entries directory_id="19" orderby="date" sort="desc" limit="3000" disable="relationships|categories|category_fields|member_data|pagination"}

        {!--<!-- [STORE PHP VALUE IN EE VARIABLE TO USE LATER] -->--}
        <?php 
            // break apart file name, grab first word
            $str = '{title}'; 
            $arr1 = preg_split("/[_,\- ]+/", $str);
            //echo $arr1[0]."\n"; 
        ?>
        {stash:batch_value}<?php echo $arr1[0]."\n"; ?>{/stash:batch_value}

        {stash:batch_number}{exp:stringer:striptags words="1"}{title}{/exp:stringer:striptags}{/stash:batch_number}



        {stash:entry_title}{title}{/stash:entry_title}
        {stash:count}{count}{/stash:count}
        {stash:url}{file_url}{/stash:url}

    {/exp:file:entries}

{/exp:stash:set_list}

{exp:stash:get_list name="my_entries" match="#(?i){freebie_3}#" against="batch_number"}

    <div class="item {batch_number} field-id-{count}">
        <p><span>{entry_title}</span> // {batch_value} - <a href="{url}" title="Download {entry_title}" download>Download</a></p>
    </div>

        {if no_results}
            {exp:stash:set name="my_message"}No results{/exp:stash:set}
        {/if}

{/exp:stash:get_list}        

This is working almost perfectly.

But now, when I have php parse set to input, I get the full file name (doesn't break it apart and give me the first word). When I have php parse set to output, I get this error:

NOTICE UNDEFINED VARIABLE: ARR1 ee/legacy/libraries/Functions.php(682) : eval()'d code, line 46
Severity: E_NOTICE

I've tried using stringer to get the first word, but it doesn't break on underscores, and I need that :/

Answer 1

I got some feedback from the developer of Stash (Thanks again, Mark!) on this via Twitter, and he advised not to do it this way.

So I went back to the drawing board. I ended up having our programmer help me out and write some custom php to perform a search in the file directory, and return the results in my template, rather than use any EE tags.

Basically, the php converts all of the file names to lower-case. Then searches the filenames to match against the key-term entered. Removes the server path from the file so we have a clean URL and display file name. Then writes the results to the template.