feat: support uploading pdf in document knowledgebase

2025-06-26 02:41:19 +05:30
parent c42dd8a23e
commit ea3a84c930
14 changed files with 864 additions and 35 deletions
@@ -94,3 +94,13 @@ yarn-debug.log*
 .vscode
 .claude/settings.local.json
 .cursor
+
+bin/
+local/
+cache/
+gems/
+specifications/
+extensions/
+build_info/
+bin
+.pnpm-store/
@@ -15,6 +15,14 @@ class CaptainDocument extends ApiClient {
      },
    });
  }
+
+  uploadPdf(formData) {
+    return axios.post(`${this.url}/upload_pdf`, formData, {
+      headers: {
+        'Content-Type': 'multipart/form-data',
+      },
+    });
+  }
 }

 export default new CaptainDocument();
@@ -18,12 +18,29 @@ const i18nKey = 'CAPTAIN.DOCUMENTS.CREATE';

 const handleSubmit = async newDocument => {
  try {
-    await store.dispatch('captainDocuments/create', newDocument);
-    useAlert(t(`${i18nKey}.SUCCESS_MESSAGE`));
+    if (newDocument.type === 'pdf') {
+      // Handle PDF upload
+      const formData = new FormData();
+      formData.append('pdf_document', newDocument.pdf_document);
+      formData.append('assistant_id', newDocument.assistant_id);
+
+      await store.dispatch('captainDocuments/uploadPdf', formData);
+      useAlert('PDF uploaded successfully! Processing will begin shortly.');
+    } else {
+      // Handle URL creation
+      await store.dispatch('captainDocuments/create', {
+        document: {
+          external_link: newDocument.external_link,
+          assistant_id: newDocument.assistant_id,
+        },
+      });
+      useAlert('Document created successfully!');
+    }
    dialogRef.value.close();
  } catch (error) {
    const errorMessage =
-      error?.response?.message || t(`${i18nKey}.ERROR_MESSAGE`);
+      error?.response?.data?.message ||
+      'Failed to create document. Please try again.';
    useAlert(errorMessage);
  }
 };
@@ -42,8 +59,8 @@ defineExpose({ dialogRef });
 <template>
  <Dialog
    ref="dialogRef"
-    :title="$t(`${i18nKey}.TITLE`)"
-    :description="$t('CAPTAIN.DOCUMENTS.FORM_DESCRIPTION')"
+    title="Create New Document"
+    description="Add a new document from a website URL or by uploading a PDF file"
    :show-cancel-button="false"
    :show-confirm-button="false"
    @close="handleClose"
@@ -1,5 +1,5 @@
 <script setup>
-import { reactive, computed } from 'vue';
+import { reactive, computed, ref } from 'vue';
 import { useI18n } from 'vue-i18n';
 import { useVuelidate } from '@vuelidate/core';
 import { required, minLength, url } from '@vuelidate/validators';
@@ -8,6 +8,7 @@ import { useMapGetter } from 'dashboard/composables/store';
 import Input from 'dashboard/components-next/input/Input.vue';
 import Button from 'dashboard/components-next/button/Button.vue';
 import ComboBox from 'dashboard/components-next/combobox/ComboBox.vue';
+import FileUpload from 'dashboard/components-next/fileUpload/FileUpload.vue';

 const emit = defineEmits(['submit', 'cancel']);

@@ -19,16 +20,26 @@ const formState = {
 };

 const initialState = {
-  name: '',
+  url: '',
  assistantId: null,
+  selectedFile: null,
+  sourceType: 'url', // 'url' or 'pdf'
 };

 const state = reactive({ ...initialState });
+const uploadError = ref('');

-const validationRules = {
-  url: { required, url, minLength: minLength(1) },
-  assistantId: { required },
-};
+const validationRules = computed(() => {
+  const rules = {
+    assistantId: { required },
+  };
+
+  if (state.sourceType === 'url') {
+    rules.url = { required, url, minLength: minLength(1) };
+  }
+
+  return rules;
+});

 const assistantList = computed(() =>
  formState.assistants.value.map(assistant => ({
@@ -41,70 +52,252 @@ const v$ = useVuelidate(validationRules, state);

 const isLoading = computed(() => formState.uiFlags.value.creatingItem);

-const getErrorMessage = (field, errorKey) => {
-  return v$.value[field].$error
-    ? t(`CAPTAIN.DOCUMENTS.FORM.${errorKey}.ERROR`)
-    : '';
+const getErrorMessage = (field, errorMessage) => {
+  return v$.value[field]?.$error ? errorMessage : '';
 };

 const formErrors = computed(() => ({
-  url: getErrorMessage('url', 'URL'),
-  assistantId: getErrorMessage('assistantId', 'ASSISTANT'),
+  url: getErrorMessage('url', 'Please enter a valid URL'),
+  assistantId: getErrorMessage('assistantId', 'Please select an assistant'),
 }));

 const handleCancel = () => emit('cancel');

-const prepareDocumentDetails = () => ({
-  external_link: state.url,
-  assistant_id: state.assistantId,
-});
+const prepareDocumentDetails = () => {
+  if (state.sourceType === 'pdf') {
+    return {
+      type: 'pdf',
+      pdf_document: state.selectedFile,
+      assistant_id: state.assistantId,
+    };
+  }
+
+  return {
+    type: 'url',
+    external_link: state.url,
+    assistant_id: state.assistantId,
+  };
+};

 const handleSubmit = async () => {
+  uploadError.value = '';
+
+  // Validate form
  const isFormValid = await v$.value.$validate();
  if (!isFormValid) {
    return;
  }

+  // For PDF uploads, check if file is selected
+  if (state.sourceType === 'pdf' && !state.selectedFile) {
+    uploadError.value = 'Please select a PDF file to upload.';
+    return;
+  }
+
  emit('submit', prepareDocumentDetails());
 };
+
+// PDF Upload handlers
+const handleFileSelected = file => {
+  state.selectedFile = file;
+  uploadError.value = '';
+};
+
+const handleFileError = error => {
+  uploadError.value = error;
+  state.selectedFile = null;
+};
+
+const clearSelectedFile = () => {
+  state.selectedFile = null;
+  uploadError.value = '';
+};
+
+const handleSourceTypeChange = type => {
+  state.sourceType = type;
+  uploadError.value = '';
+  state.selectedFile = null;
+  state.url = '';
+};
 </script>

 <template>
  <form class="flex flex-col gap-4" @submit.prevent="handleSubmit">
-    <Input
-      v-model="state.url"
-      :label="t('CAPTAIN.DOCUMENTS.FORM.URL.LABEL')"
-      :placeholder="t('CAPTAIN.DOCUMENTS.FORM.URL.PLACEHOLDER')"
-      :message="formErrors.url"
-      :message-type="formErrors.url ? 'error' : 'info'"
-    />
+    <!-- Source Type Selection -->
+    <div class="flex flex-col gap-2">
+      <label class="mb-0.5 text-sm font-medium text-n-slate-12">
+        Choose Source Type
+      </label>
+      <div class="flex gap-2">
+        <button
+          type="button"
+          class="flex-1 p-3 rounded-lg border border-n-slate-6 hover:border-n-slate-7 transition-colors"
+          :class="{
+            'bg-primary-50 border-primary-500 text-primary-700':
+              state.sourceType === 'url',
+            'bg-n-alpha-2 text-n-slate-10': state.sourceType !== 'url',
+          }"
+          @click="handleSourceTypeChange('url')"
+        >
+          <div class="flex flex-col items-center gap-2">
+            <svg
+              width="20"
+              height="20"
+              viewBox="0 0 24 24"
+              fill="none"
+              xmlns="http://www.w3.org/2000/svg"
+            >
+              <path
+                d="M3.9 12C3.9 10.29 5.29 8.9 7 8.9H11V7H7C4.24 7 2 9.24 2 12S4.24 17 7 17H11V15.1H7C5.29 15.1 3.9 13.71 3.9 12ZM8 13H16V11H8V13ZM17 7H13V8.9H17C18.71 8.9 20.1 10.29 20.1 12S18.71 15.1 17 15.1H13V17H17C19.76 17 22 14.76 22 12S19.76 7 17 7Z"
+                fill="currentColor"
+              />
+            </svg>
+            <span class="text-sm font-medium">Website URL</span>
+          </div>
+        </button>
+        <button
+          type="button"
+          class="flex-1 p-3 rounded-lg border border-n-slate-6 hover:border-n-slate-7 transition-colors"
+          :class="{
+            'bg-primary-50 border-primary-500 text-primary-700':
+              state.sourceType === 'pdf',
+            'bg-n-alpha-2 text-n-slate-10': state.sourceType !== 'pdf',
+          }"
+          @click="handleSourceTypeChange('pdf')"
+        >
+          <div class="flex flex-col items-center gap-2">
+            <svg
+              width="20"
+              height="20"
+              viewBox="0 0 24 24"
+              fill="none"
+              xmlns="http://www.w3.org/2000/svg"
+            >
+              <path
+                d="M14,2H6A2,2 0 0,0 4,4V20A2,2 0 0,0 6,22H18A2,2 0 0,0 20,20V8L14,2M18,20H6V4H13V9H18V20Z"
+                fill="currentColor"
+              />
+            </svg>
+            <span class="text-sm font-medium">PDF Upload</span>
+          </div>
+        </button>
+      </div>
+    </div>
+
+    <!-- URL Input (when URL is selected) -->
+    <div v-if="state.sourceType === 'url'">
+      <Input
+        v-model="state.url"
+        label="Website URL"
+        placeholder="Enter the website URL to crawl"
+        :message="formErrors.url"
+        :message-type="formErrors.url ? 'error' : 'info'"
+      />
+    </div>
+
+    <!-- PDF Upload (when PDF is selected) -->
+    <div v-if="state.sourceType === 'pdf'" class="flex flex-col gap-2">
+      <label class="mb-0.5 text-sm font-medium text-n-slate-12">
+        Upload PDF Document
+      </label>
+
+      <!-- File upload area -->
+      <div v-if="!state.selectedFile">
+        <FileUpload
+          accept="application/pdf"
+          :max-size-m-b="10"
+          placeholder="Select a PDF file"
+          upload-text="Click to select PDF or drag and drop"
+          drag-text="Drop PDF file here"
+          @file-selected="handleFileSelected"
+          @file-error="handleFileError"
+        />
+      </div>
+
+      <!-- Selected file display -->
+      <div v-else class="p-3 bg-n-alpha-2 rounded-lg border border-n-slate-6">
+        <div class="flex items-center justify-between">
+          <div class="flex items-center gap-3">
+            <svg
+              width="20"
+              height="20"
+              viewBox="0 0 24 24"
+              fill="none"
+              xmlns="http://www.w3.org/2000/svg"
+              class="text-red-600"
+            >
+              <path
+                d="M14,2H6A2,2 0 0,0 4,4V20A2,2 0 0,0 6,22H18A2,2 0 0,0 20,20V8L14,2M18,20H6V4H13V9H18V20Z"
+                fill="currentColor"
+              />
+            </svg>
+            <div>
+              <p class="text-sm font-medium text-n-slate-12">
+                {{ state.selectedFile.name }}
+              </p>
+              <p class="text-xs text-n-slate-8">
+                {{
+                  Math.round((state.selectedFile.size / 1024 / 1024) * 100) /
+                  100
+                }}
+                MB
+              </p>
+            </div>
+          </div>
+          <button
+            type="button"
+            class="p-1 rounded hover:bg-n-alpha-3 text-n-slate-8 hover:text-n-slate-10"
+            @click="clearSelectedFile"
+          >
+            <svg
+              width="16"
+              height="16"
+              viewBox="0 0 24 24"
+              fill="none"
+              xmlns="http://www.w3.org/2000/svg"
+            >
+              <path
+                d="M19,6.41L17.59,5L12,10.59L6.41,5L5,6.41L10.59,12L5,17.59L6.41,19L12,13.41L17.59,19L19,17.59L13.41,12L19,6.41Z"
+                fill="currentColor"
+              />
+            </svg>
+          </button>
+        </div>
+      </div>
+
+      <!-- Upload error -->
+      <p v-if="uploadError" class="text-sm text-red-600">{{ uploadError }}</p>
+    </div>
+
+    <!-- Assistant Selection -->
    <div class="flex flex-col gap-1">
      <label for="assistant" class="mb-0.5 text-sm font-medium text-n-slate-12">
-        {{ t('CAPTAIN.DOCUMENTS.FORM.ASSISTANT.LABEL') }}
+        Select Assistant
      </label>
      <ComboBox
        id="assistant"
        v-model="state.assistantId"
        :options="assistantList"
        :has-error="!!formErrors.assistantId"
-        :placeholder="t('CAPTAIN.DOCUMENTS.FORM.ASSISTANT.PLACEHOLDER')"
+        placeholder="Choose an assistant"
        class="[&>div>button]:bg-n-alpha-black2 [&>div>button:not(.focused)]:dark:outline-n-weak [&>div>button:not(.focused)]:hover:!outline-n-slate-6"
        :message="formErrors.assistantId"
      />
    </div>

+    <!-- Action Buttons -->
    <div class="flex items-center justify-between w-full gap-3">
      <Button
        type="button"
        variant="faded"
        color="slate"
-        :label="t('CAPTAIN.FORM.CANCEL')"
+        label="Cancel"
        class="w-full bg-n-alpha-2 n-blue-text hover:bg-n-alpha-3"
        @click="handleCancel"
      />
      <Button
        type="submit"
-        :label="t('CAPTAIN.FORM.CREATE')"
+        label="Create Document"
        class="w-full"
        :is-loading="isLoading"
        :disabled="isLoading"
@@ -0,0 +1,201 @@
+<script setup>
+import { ref, computed } from 'vue';
+import { useI18n } from 'vue-i18n';
+
+const props = defineProps({
+  accept: {
+    type: String,
+    default: '*',
+  },
+  multiple: {
+    type: Boolean,
+    default: false,
+  },
+  maxSizeMB: {
+    type: Number,
+    default: 10,
+  },
+  disabled: {
+    type: Boolean,
+    default: false,
+  },
+  placeholder: {
+    type: String,
+    default: '',
+  },
+  uploadText: {
+    type: String,
+    default: '',
+  },
+  dragText: {
+    type: String,
+    default: '',
+  },
+});
+
+const emit = defineEmits(['file-selected', 'file-error']);
+
+const { t } = useI18n();
+const fileInput = ref(null);
+const isDragOver = ref(false);
+
+const acceptTypes = computed(() => props.accept);
+const maxSize = computed(() => props.maxSizeMB * 1024 * 1024);
+
+const handleFileClick = () => {
+  if (!props.disabled) {
+    fileInput.value?.click();
+  }
+};
+
+const validateFile = file => {
+  // Check file size
+  if (file.size > maxSize.value) {
+    const errorMsg = `File "${file.name}" is too large. Maximum size is ${props.maxSizeMB}MB.`;
+    emit('file-error', errorMsg);
+    return false;
+  }
+
+  // Check file type if specified
+  if (
+    props.accept !== '*' &&
+    !file.type.match(new RegExp(props.accept.replace('*', '.*')))
+  ) {
+    const errorMsg = `File type not supported. Only ${props.accept} files are allowed.`;
+    emit('file-error', errorMsg);
+    return false;
+  }
+
+  return true;
+};
+
+const handleFileChange = event => {
+  const files = event.target.files;
+  if (files && files.length > 0) {
+    const file = files[0];
+    if (validateFile(file)) {
+      emit('file-selected', file);
+    }
+  }
+  // Reset input value to allow selecting the same file again
+  event.target.value = '';
+};
+
+const handleDragOver = event => {
+  event.preventDefault();
+  if (!props.disabled) {
+    isDragOver.value = true;
+  }
+};
+
+const handleDragLeave = event => {
+  event.preventDefault();
+  // Only set to false if we're leaving the drop zone completely
+  if (!event.currentTarget.contains(event.relatedTarget)) {
+    isDragOver.value = false;
+  }
+};
+
+const handleDrop = event => {
+  event.preventDefault();
+  isDragOver.value = false;
+
+  if (props.disabled) return;
+
+  const files = event.dataTransfer.files;
+  if (files && files.length > 0) {
+    const file = files[0];
+    if (validateFile(file)) {
+      emit('file-selected', file);
+    }
+  }
+};
+
+const displayText = computed(() => {
+  if (isDragOver.value && props.dragText) {
+    return props.dragText;
+  }
+  return props.uploadText || props.placeholder;
+});
+</script>
+
+<template>
+  <div
+    class="file-upload-zone"
+    :class="{
+      'drag-over': isDragOver,
+      disabled: disabled,
+    }"
+    @dragover="handleDragOver"
+    @dragleave="handleDragLeave"
+    @drop="handleDrop"
+    @click="handleFileClick"
+  >
+    <input
+      ref="fileInput"
+      type="file"
+      :accept="acceptTypes"
+      :multiple="multiple"
+      :disabled="disabled"
+      class="hidden"
+      @change="handleFileChange"
+    />
+
+    <div class="upload-content">
+      <div class="upload-icon">
+        <svg
+          width="24"
+          height="24"
+          viewBox="0 0 24 24"
+          fill="none"
+          xmlns="http://www.w3.org/2000/svg"
+          class="text-n-slate-8"
+        >
+          <path
+            d="M14,2H6A2,2 0 0,0 4,4V20A2,2 0 0,0 6,22H18A2,2 0 0,0 20,20V8L14,2M18,20H6V4H13V9H18V20Z"
+            fill="currentColor"
+          />
+        </svg>
+      </div>
+      <div class="upload-text">
+        <p class="text-sm font-medium text-n-slate-11">
+          {{ displayText }}
+        </p>
+        <p class="text-xs text-n-slate-8 mt-1">
+          Maximum file size: {{ maxSizeMB }}MB
+        </p>
+      </div>
+    </div>
+  </div>
+</template>
+
+<style scoped>
+.file-upload-zone {
+  @apply border-2 border-dashed border-n-slate-6 rounded-lg p-6;
+  @apply bg-n-alpha-2 hover:bg-n-alpha-3 cursor-pointer;
+  @apply transition-all duration-200 ease-in-out;
+  @apply flex flex-col items-center justify-center text-center;
+  min-height: 120px;
+}
+
+.file-upload-zone:hover {
+  @apply border-n-slate-7;
+}
+
+.file-upload-zone.drag-over {
+  @apply border-primary-500 bg-primary-50/50;
+}
+
+.file-upload-zone.disabled {
+  @apply opacity-50 cursor-not-allowed;
+  @apply hover:bg-n-alpha-2 hover:border-n-slate-6;
+}
+
+.upload-content {
+  @apply flex flex-col items-center gap-3;
+}
+
+.upload-icon {
+  @apply p-2 rounded-full bg-n-alpha-3;
+}
+</style>
@@ -1,7 +1,24 @@
 import CaptainDocumentAPI from 'dashboard/api/captain/document';
 import { createStore } from './storeFactory';

+const actions = mutationTypes => ({
+  async uploadPdf({ commit }, formData) {
+    commit(mutationTypes.SET_UI_FLAG, { creatingItem: true });
+    try {
+      const response = await CaptainDocumentAPI.uploadPdf(formData);
+      const { data } = response;
+      commit(mutationTypes.ADD, data.document);
+      commit(mutationTypes.SET_UI_FLAG, { creatingItem: false });
+      return Promise.resolve(data);
+    } catch (error) {
+      commit(mutationTypes.SET_UI_FLAG, { creatingItem: false });
+      return Promise.reject(error);
+    }
+  },
+});
+
 export default createStore({
  name: 'CaptainDocument',
  API: CaptainDocumentAPI,
+  actions,
 });
@@ -63,7 +63,11 @@ Rails.application.routes.draw do
            resources :copilot_threads, only: [:index, :create] do
              resources :copilot_messages, only: [:index, :create]
            end
-            resources :documents, only: [:index, :show, :create, :destroy]
+            resources :documents, only: [:index, :show, :create, :destroy] do
+              collection do
+                post :upload_pdf
+              end
+            end
          end
          resources :agent_bots, only: [:index, :create, :show, :update, :destroy] do
            delete :avatar, on: :member
@@ -5,7 +5,7 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
  before_action :set_current_page, only: [:index]
  before_action :set_documents, except: [:create]
  before_action :set_document, only: [:show, :destroy]
-  before_action :set_assistant, only: [:create]
+  before_action :set_assistant, only: [:create, :upload_pdf]
  RESULTS_PER_PAGE = 25

  def index
@@ -27,6 +27,35 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
    render_could_not_create_error(e.message)
  end

+  def upload_pdf
+    return render_could_not_create_error('Missing Assistant') if @assistant.nil?
+    return render_could_not_create_error('No PDF file provided') unless pdf_file_present?
+    
+    # Validate PDF file
+    return render_could_not_create_error('Invalid file type. Only PDF files are allowed.') unless valid_pdf_file?
+    return render_could_not_create_error('File size too large. Maximum size is 10MB.') unless valid_pdf_size?
+    
+    # Upload PDF to storage and get URL
+    blob = create_pdf_blob
+    pdf_url = Rails.application.routes.url_helpers.rails_blob_url(blob, host: ENV.fetch('FRONTEND_URL', 'http://localhost:3000'))
+    
+    # Create document with PDF URL
+    @document = @assistant.documents.build(
+      name: pdf_params[:pdf_document].original_filename.gsub('.pdf', ''),
+      external_link: pdf_url
+    )
+    @document.save!
+    
+    render json: { 
+      document: @document.as_json(only: [:id, :name, :status, :created_at]),
+      message: 'PDF uploaded successfully. Processing will begin shortly.'
+    }
+  rescue Captain::Document::LimitExceededError => e
+    render_could_not_create_error(e.message)
+  rescue StandardError => e
+    render_could_not_create_error("PDF upload failed: #{e.message}")
+  end
+
  def destroy
    @document.destroy
    head :no_content
@@ -43,7 +72,8 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
  end

  def set_assistant
-    @assistant = Current.account.captain_assistants.find_by(id: document_params[:assistant_id])
+    assistant_id = action_name == 'upload_pdf' ? pdf_params[:assistant_id] : document_params[:assistant_id]
+    @assistant = Current.account.captain_assistants.find_by(id: assistant_id)
  end

  def set_current_page
@@ -57,4 +87,32 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
  def document_params
    params.require(:document).permit(:name, :external_link, :assistant_id)
  end
+
+  def pdf_params
+    params.permit(:pdf_document, :assistant_id)
+  end
+
+  def pdf_file_present?
+    pdf_params[:pdf_document].present?
+  end
+
+  def valid_pdf_file?
+    return false unless pdf_params[:pdf_document].respond_to?(:content_type)
+    
+    pdf_params[:pdf_document].content_type == 'application/pdf'
+  end
+
+  def valid_pdf_size?
+    return false unless pdf_params[:pdf_document].respond_to?(:size)
+    
+    pdf_params[:pdf_document].size <= 10.megabytes
+  end
+
+  def create_pdf_blob
+    ActiveStorage::Blob.create_and_upload!(
+      io: pdf_params[:pdf_document].tempfile,
+      filename: pdf_params[:pdf_document].original_filename,
+      content_type: pdf_params[:pdf_document].content_type
+    )
+  end
 end
@@ -22,4 +22,8 @@ class Captain::AssistantPolicy < ApplicationPolicy
  def playground?
    true
  end
+
+  def upload_pdf?
+    @account_user.administrator?
+  end
 end
@@ -0,0 +1,100 @@
+require 'rails_helper'
+
+RSpec.describe Api::V1::Accounts::Captain::DocumentsController, type: :controller do
+  let(:account) { create(:account) }
+  let(:user) { create(:user, account: account) }
+  let(:assistant) { create(:captain_assistant, account: account) }
+
+  before do
+    sign_in(user)
+  end
+
+  describe 'POST #upload_pdf' do
+    let(:pdf_file) do
+      fixture_file_upload(
+        Rails.root.join('spec', 'fixtures', 'files', 'sample.pdf'),
+        'application/pdf'
+      )
+    end
+
+    let(:valid_params) do
+      {
+        account_id: account.id,
+        pdf_document: pdf_file,
+        assistant_id: assistant.id
+      }
+    end
+
+    context 'with valid parameters' do
+      it 'creates a new document' do
+        expect do
+          post :upload_pdf, params: valid_params
+        end.to change(Captain::Document, :count).by(1)
+      end
+
+      it 'returns success response' do
+        post :upload_pdf, params: valid_params
+        expect(response).to have_http_status(:ok)
+        expect(JSON.parse(response.body)).to have_key('document')
+        expect(JSON.parse(response.body)).to have_key('message')
+      end
+
+      it 'sets document name from filename' do
+        post :upload_pdf, params: valid_params
+        document = Captain::Document.last
+        expect(document.name).to eq('sample')
+      end
+
+      it 'sets document status to in_progress' do
+        post :upload_pdf, params: valid_params
+        document = Captain::Document.last
+        expect(document.status).to eq('in_progress')
+      end
+    end
+
+    context 'with invalid parameters' do
+      it 'returns error when assistant is missing' do
+        params = valid_params.except(:assistant_id)
+        post :upload_pdf, params: params
+        expect(response).to have_http_status(:unprocessable_entity)
+      end
+
+      it 'returns error when PDF file is missing' do
+        params = valid_params.except(:pdf_document)
+        post :upload_pdf, params: params
+        expect(response).to have_http_status(:unprocessable_entity)
+      end
+
+      it 'returns error when file is not a PDF' do
+        txt_file = fixture_file_upload(
+          Rails.root.join('spec', 'fixtures', 'files', 'sample.txt'),
+          'text/plain'
+        )
+        params = valid_params.merge(pdf_document: txt_file)
+        post :upload_pdf, params: params
+        expect(response).to have_http_status(:unprocessable_entity)
+      end
+
+      it 'returns error when file is too large' do
+        # Mock large file size
+        allow_any_instance_of(ActionDispatch::Http::UploadedFile).to receive(:size).and_return(11.megabytes)
+        post :upload_pdf, params: valid_params
+        expect(response).to have_http_status(:unprocessable_entity)
+      end
+    end
+
+    context 'when document limit is exceeded' do
+      before do
+        allow_any_instance_of(Captain::Document).to receive(:save!).and_raise(
+          Captain::Document::LimitExceededError, 'Document limit exceeded'
+        )
+      end
+
+      it 'returns limit exceeded error' do
+        post :upload_pdf, params: valid_params
+        expect(response).to have_http_status(:unprocessable_entity)
+        expect(JSON.parse(response.body)['message']).to include('Document limit exceeded')
+      end
+    end
+  end
+end
@@ -288,4 +288,69 @@ RSpec.describe 'Api::V1::Accounts::Captain::Documents', type: :request do
      end
    end
  end
+
+  describe 'POST /api/v1/accounts/:account_id/captain/documents/upload_pdf' do
+    let(:pdf_file) { fixture_file_upload('spec/fixtures/files/sample.pdf', 'application/pdf') }
+    let(:valid_pdf_params) do
+      {
+        pdf_document: pdf_file,
+        assistant_id: assistant.id
+      }
+    end
+
+    context 'when it is an un-authenticated user' do
+      before do
+        post "/api/v1/accounts/#{account.id}/captain/documents/upload_pdf",
+             params: valid_pdf_params
+      end
+
+      it 'returns unauthorized status' do
+        expect(response).to have_http_status(:unauthorized)
+      end
+    end
+
+    context 'when it is an agent' do
+      it 'returns unauthorized' do
+        post "/api/v1/accounts/#{account.id}/captain/documents/upload_pdf",
+             params: valid_pdf_params,
+             headers: agent.create_new_auth_token
+
+        expect(response).to have_http_status(:unauthorized)
+      end
+    end
+
+    context 'when it is an admin' do
+      context 'with valid PDF file' do
+        it 'uploads PDF and creates document' do
+          expect do
+            post "/api/v1/accounts/#{account.id}/captain/documents/upload_pdf",
+                 params: valid_pdf_params,
+                 headers: admin.create_new_auth_token
+          end.to change(Captain::Document, :count).by(1)
+        end
+
+        it 'returns success status with document data' do
+          post "/api/v1/accounts/#{account.id}/captain/documents/upload_pdf",
+               params: valid_pdf_params,
+               headers: admin.create_new_auth_token, as: :json
+
+          expect(response).to have_http_status(:success)
+          expect(json_response[:document]).to be_present
+          expect(json_response[:message]).to eq('PDF uploaded successfully. Processing will begin shortly.')
+        end
+      end
+
+      context 'without PDF file' do
+        before do
+          post "/api/v1/accounts/#{account.id}/captain/documents/upload_pdf",
+               params: { assistant_id: assistant.id },
+               headers: admin.create_new_auth_token
+        end
+
+        it 'returns unprocessable entity status' do
+          expect(response).to have_http_status(:unprocessable_entity)
+        end
+      end
+    end
+  end
 end
@@ -0,0 +1,53 @@
+%PDF-1.4
+1 0 obj
+<<
+/Type /Catalog
+/Pages 2 0 R
+>>
+endobj
+
+2 0 obj
+<<
+/Type /Pages
+/Kids [3 0 R]
+/Count 1
+>>
+endobj
+
+3 0 obj
+<<
+/Type /Page
+/Parent 2 0 R
+/MediaBox [0 0 612 792]
+/Contents 4 0 R
+>>
+endobj
+
+4 0 obj
+<<
+/Length 44
+>>
+stream
+BT
+/F1 12 Tf
+72 720 Td
+(Sample PDF for testing) Tj
+ET
+endstream
+endobj
+
+xref
+0 5
+0000000000 65535 f 
+0000000009 00000 n 
+0000000058 00000 n 
+0000000115 00000 n 
+0000000202 00000 n 
+trailer
+<<
+/Size 5
+/Root 1 0 R
+>>
+startxref
+294
+%%EOF
@@ -0,0 +1,2 @@
+This is a sample text file for testing purposes.
+It contains some basic text content that can be used in tests.
@@ -0,0 +1,97 @@
+require 'rails_helper'
+
+RSpec.describe Captain::Documents::CrawlJob, type: :job do
+  include ActiveJob::TestHelper
+
+  let(:account) { create(:account) }
+  let(:assistant) { create(:captain_assistant, account: account) }
+  let(:document) { create(:captain_document, assistant: assistant, account: account) }
+
+  describe '#perform' do
+    context 'when Firecrawl API key is configured' do
+      before do
+        create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test_api_key')
+      end
+
+      it 'calls perform_firecrawl_crawl' do
+        expect_any_instance_of(described_class).to receive(:perform_firecrawl_crawl).with(document)
+        described_class.perform_now(document)
+      end
+
+      it 'handles PDF URLs correctly' do
+        pdf_document = create(:captain_document, 
+          assistant: assistant, 
+          account: account,
+          external_link: 'https://example.com/document.pdf'
+        )
+        
+        expect(Captain::Tools::FirecrawlService).to receive_message_chain(:new, :perform)
+          .with(pdf_document.external_link, anything, anything)
+        
+        described_class.perform_now(pdf_document)
+      end
+    end
+
+    context 'when Firecrawl API key is not configured' do
+      before do
+        InstallationConfig.find_by(name: 'CAPTAIN_FIRECRAWL_API_KEY')&.destroy
+      end
+
+      it 'calls perform_simple_crawl' do
+        expect_any_instance_of(described_class).to receive(:perform_simple_crawl).with(document)
+        described_class.perform_now(document)
+      end
+    end
+
+    context 'webhook URL generation' do
+      before do
+        create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test_api_key')
+      end
+
+      it 'generates correct webhook URL with assistant_id and token' do
+        job = described_class.new
+        webhook_url = job.send(:firecrawl_webhook_url, document)
+        
+        expect(webhook_url).to include("assistant_id=#{document.assistant_id}")
+        expect(webhook_url).to include('token=')
+        expect(webhook_url).to include('/enterprise/webhooks/firecrawl')
+      end
+    end
+
+    context 'crawl limit calculation' do
+      before do
+        create(:installation_config, name: 'CAPTAIN_FIRECRAWL_API_KEY', value: 'test_api_key')
+      end
+
+      it 'respects account usage limits' do
+        allow(document.account).to receive(:usage_limits).and_return({
+          captain: {
+            documents: {
+              current_available: 5
+            }
+          }
+        })
+
+        expect(Captain::Tools::FirecrawlService).to receive_message_chain(:new, :perform)
+          .with(anything, anything, 5)
+        
+        described_class.perform_now(document)
+      end
+
+      it 'caps crawl limit at 500' do
+        allow(document.account).to receive(:usage_limits).and_return({
+          captain: {
+            documents: {
+              current_available: 1000
+            }
+          }
+        })
+
+        expect(Captain::Tools::FirecrawlService).to receive_message_chain(:new, :perform)
+          .with(anything, anything, 500)
+        
+        described_class.perform_now(document)
+      end
+    end
+  end
+end