Day 4 – Storing and Displaying GPT-Extracted Document Fields #LaravelGPT #SmartDocs #DocumentParsing #AIValidation #LegalAI #LaravelAI

Today we’ll store the structured fields extracted by GPT and display them in a readable format. We’ll also prepare the UI to allow admins to review and correct the AI output before finalizing.

Table of Contents

🧩 Step 1: Add a JSON Column to Store Fields

Update your documents table:

php artisan make:migration add_extracted_fields_to_documents_table

In the migration:

public function up()
{
    Schema::table('documents', function (Blueprint $table) {
        $table->json('extracted_fields')->nullable();
    });
}

Then:

php artisan migrate

💾 Step 2: Save GPT Fields to This Column

Update your controller:

Document::create([
    'title' => $request->title,
    'type' => $request->type,
    'filename' => $filename,
    'user_id' => auth()->id(),
    'extracted_text' => $text,
    'extracted_fields' => $fields, // 👈 added
]);

📄 Step 3: Show Fields in the View

In resources/views/documents/show.blade.php:

<h2>Extracted Fields</h2>
@if($document->extracted_fields)
    <ul>
        @foreach(json_decode($document->extracted_fields, true) as $key => $value)
            <li><strong>{{ ucfirst(str_replace('_', ' ', $key)) }}:</strong> {{ $value }}</li>
        @endforeach
    </ul>
@else
    <p>No extracted fields yet.</p>
@endif

✏️ Step 4: Add Editable Fields (Optional)

To make the fields editable, turn each into a form input:

<form method="POST" action="{{ route('documents.updateFields', $document->id) }}">
    @csrf
    @method('PUT')
    @foreach(json_decode($document->extracted_fields, true) as $key => $value)
        <label>{{ ucfirst(str_replace('_', ' ', $key)) }}</label>
        <input type="text" name="fields[{{ $key }}]" value="{{ $value }}">
    @endforeach
    <button type="submit">Save Changes</button>
</form>

Then define a route and controller method:

Route::put('/documents/{document}/fields', [DocumentController::class, 'updateFields'])->name('documents.updateFields');

In controller:

public function updateFields(Request $request, Document $document)
{
    $document->update([
        'extracted_fields' => $request->input('fields'),
    ]);

    return back()->with('success', 'Fields updated.');
}

✅ Summary

✅ Today you:

Saved GPT-extracted fields to the DB
Displayed them nicely in the UI
Made them editable for admin review

✅ Up next (Day 5): We’ll build a validation checklist to flag missing fields, invalid data, and risky clauses using GPT prompts.

Post Views: 4