Merge pull request #118 from scub-france/feature/configurable-max-file-size

feat: make document max size configurable via MAX_FILE_SIZE_MB

.env.example

@@ -15,6 +15,9 @@
 # Max parallel analysis jobs (default: 3)
 # MAX_CONCURRENT_ANALYSES=3
 
+# Max upload file size in MB (default: 50, 0 = unlimited)
+# MAX_FILE_SIZE_MB=50
+
 # Max pages per PDF (default: 0 = unlimited). Set to 20 for HF Spaces.
 # MAX_PAGE_COUNT=0
 

document-parser/api/documents.py

@@ -8,6 +8,7 @@ from fastapi import APIRouter, HTTPException, Query, UploadFile
 from fastapi.responses import Response
 
 from api.schemas import DocumentResponse
+from infra.settings import settings
 from services import document_service
 
 logger = logging.getLogger(__name__)
@@ -34,16 +35,18 @@ async def upload(file: UploadFile) -> DocumentResponse:
         raise HTTPException(status_code=400, detail="No filename provided")
 
     # Reject early if Content-Length exceeds limit (before reading body)
-    if file.size and file.size > document_service.MAX_FILE_SIZE:
-        raise HTTPException(status_code=413, detail="File too large (max 5 MB)")
+    _max = document_service.MAX_FILE_SIZE
+    _detail = f"File too large (max {settings.max_file_size_mb} MB)"
+    if _max > 0 and file.size and file.size > _max:
+        raise HTTPException(status_code=413, detail=_detail)
 
     # Read in chunks to avoid holding the full upload in a single allocation
     chunks: list[bytes] = []
     total = 0
     while chunk := await file.read(_READ_CHUNK_SIZE):
         total += len(chunk)
-        if total > document_service.MAX_FILE_SIZE:
-            raise HTTPException(status_code=413, detail="File too large (max 5 MB)")
+        if _max > 0 and total > _max:
+            raise HTTPException(status_code=413, detail=_detail)
         chunks.append(chunk)
     content = b"".join(chunks)
 

document-parser/infra/settings.py

@@ -20,6 +20,7 @@ class Settings:
     default_table_mode: str = "accurate"  # "accurate" or "fast"
     max_page_count: int = 0  # 0 = unlimited (upload validation)
     max_file_size: int = 0  # 0 = unlimited (Docling-level, bytes)
+    max_file_size_mb: int = 50  # upload limit in MB (0 = unlimited)
     batch_page_size: int = 0  # 0 = disabled, > 0 = pages per batch
     upload_dir: str = "./uploads"
     db_path: str = "./data/docling_studio.db"
@@ -43,6 +44,8 @@ class Settings:
             errors.append(f"max_page_count must be >= 0 (got {self.max_page_count})")
         if self.max_file_size < 0:
             errors.append(f"max_file_size must be >= 0 (got {self.max_file_size})")
+        if self.max_file_size_mb < 0:
+            errors.append(f"max_file_size_mb must be >= 0 (got {self.max_file_size_mb})")
         if self.batch_page_size < 0:
             errors.append(f"batch_page_size must be >= 0 (got {self.batch_page_size})")
         if self.default_table_mode not in ("accurate", "fast"):
@@ -81,6 +84,7 @@ class Settings:
             default_table_mode=os.environ.get("DEFAULT_TABLE_MODE", "accurate"),
             max_page_count=int(os.environ.get("MAX_PAGE_COUNT", "0")),
             max_file_size=int(os.environ.get("MAX_FILE_SIZE", "0")),
+            max_file_size_mb=int(os.environ.get("MAX_FILE_SIZE_MB", "50")),
             batch_page_size=int(os.environ.get("BATCH_PAGE_SIZE", "0")),
             upload_dir=os.environ.get("UPLOAD_DIR", "./uploads"),
             db_path=os.environ.get("DB_PATH", "./data/docling_studio.db"),

document-parser/main.py

@@ -122,4 +122,6 @@ async def health() -> dict[str, str | int]:
     }
     if settings.max_page_count > 0:
         result["maxPageCount"] = settings.max_page_count
+    if settings.max_file_size_mb > 0:
+        result["maxFileSizeMb"] = settings.max_file_size_mb
     return result

document-parser/services/document_service.py

@@ -16,7 +16,7 @@ from persistence import analysis_repo, document_repo
 logger = logging.getLogger(__name__)
 
 UPLOAD_DIR = settings.upload_dir
-MAX_FILE_SIZE = 5 * 1024 * 1024  # 5 MB
+MAX_FILE_SIZE = settings.max_file_size_mb * 1024 * 1024 if settings.max_file_size_mb > 0 else 0
 MAX_PAGE_COUNT = settings.max_page_count  # 0 = unlimited
 
 
@@ -32,8 +32,8 @@ async def upload(filename: str, content_type: str, file_content: bytes) -> Docum
 
     Writes the file in fixed-size chunks to keep peak memory usage low.
     """
-    if len(file_content) > MAX_FILE_SIZE:
-        raise ValueError("File too large (max 5 MB)")
+    if MAX_FILE_SIZE > 0 and len(file_content) > MAX_FILE_SIZE:
+        raise ValueError(f"File too large (max {settings.max_file_size_mb} MB)")
 
     if not file_content[:4].startswith(_PDF_MAGIC):
         raise ValueError("Invalid file: not a PDF document")

document-parser/tests/test_api_endpoints.py

@@ -33,6 +33,12 @@ class TestHealthEndpoint:
         assert "engine" in data
         assert "database" in data
 
+    def test_health_exposes_max_file_size_mb(self, client):
+        resp = client.get("/api/health")
+        data = resp.json()
+        assert "maxFileSizeMb" in data
+        assert data["maxFileSizeMb"] == 50
+
 
 class TestDocumentEndpoints:
     @patch("services.document_service.find_all", new_callable=AsyncMock)

document-parser/tests/test_settings.py

@@ -17,6 +17,7 @@ class TestSettingsDefaults:
         assert s.document_timeout == 120.0
         assert s.lock_timeout == 300
         assert s.max_page_count == 0
+        assert s.max_file_size_mb == 50
         assert s.batch_page_size == 0
         assert s.upload_dir == "./uploads"
         assert s.db_path == "./data/docling_studio.db"
@@ -68,6 +69,20 @@ class TestSettingsValidation:
         with pytest.raises(ValueError, match="max_file_size must be >= 0"):
             Settings(max_file_size=-1)
 
+    def test_negative_max_file_size_mb_rejected(self):
+        import pytest
+
+        with pytest.raises(ValueError, match="max_file_size_mb must be >= 0"):
+            Settings(max_file_size_mb=-1)
+
+    def test_zero_max_file_size_mb_accepted(self):
+        s = Settings(max_file_size_mb=0)
+        assert s.max_file_size_mb == 0
+
+    def test_positive_max_file_size_mb_accepted(self):
+        s = Settings(max_file_size_mb=100)
+        assert s.max_file_size_mb == 100
+
     def test_negative_batch_page_size_rejected(self):
         import pytest
 
@@ -129,6 +144,7 @@ class TestSettingsFromEnv:
         monkeypatch.setenv("DOCUMENT_TIMEOUT", "60.0")
         monkeypatch.setenv("LOCK_TIMEOUT", "600")
         monkeypatch.setenv("MAX_PAGE_COUNT", "20")
+        monkeypatch.setenv("MAX_FILE_SIZE_MB", "100")
         monkeypatch.setenv("BATCH_PAGE_SIZE", "15")
         monkeypatch.setenv("UPLOAD_DIR", "/data/uploads")
         monkeypatch.setenv("DB_PATH", "/data/test.db")
@@ -145,6 +161,7 @@ class TestSettingsFromEnv:
         assert s.lock_timeout == 600
         assert s.document_timeout == 60.0
         assert s.max_page_count == 20
+        assert s.max_file_size_mb == 100
         assert s.batch_page_size == 15
         assert s.upload_dir == "/data/uploads"
         assert s.db_path == "/data/test.db"

frontend/src/app/App.vue

@@ -30,7 +30,7 @@
     </header>
 
     <div v-if="showDisclaimer" class="disclaimer-banner" role="alert">
-      {{ t('disclaimer.banner') }}
+      {{ t('disclaimer.banner').replace('{n}', String(flagStore.maxFileSizeMb || 50)) }}
       <button class="disclaimer-close" @click="dismissDisclaimer" aria-label="Close">
         &times;
       </button>
@@ -52,9 +52,11 @@ import { AppSidebar } from '../shared/ui/index'
 import { useSettingsStore } from '../features/settings/store'
 import { useDocumentStore } from '../features/document/store'
 import { useFeatureFlag } from '../features/feature-flags'
+import { useFeatureFlagStore } from '../features/feature-flags/store'
 import { useI18n } from '../shared/i18n'
 
 useSettingsStore()
+const flagStore = useFeatureFlagStore()
 const { t } = useI18n()
 const router = useRouter()
 const documentStore = useDocumentStore()

frontend/src/features/document/store.ts

@@ -1,11 +1,11 @@
 import { defineStore } from 'pinia'
 import { ref } from 'vue'
 import type { Document } from '../../shared/types'
+import { useFeatureFlagStore } from '../feature-flags/store'
 import * as api from './api'
 
-const MAX_FILE_SIZE = 5 * 1024 * 1024 // 5 MB
-
 export const useDocumentStore = defineStore('document', () => {
+  const flags = useFeatureFlagStore()
   const documents = ref<Document[]>([])
   const selectedId = ref<string | null>(null)
   const uploading = ref(false)
@@ -26,8 +26,9 @@ export const useDocumentStore = defineStore('document', () => {
   }
 
   async function upload(file: File): Promise<Document> {
-    if (file.size > MAX_FILE_SIZE) {
-      error.value = 'File too large (max 5 MB)'
+    const maxMb = flags.maxFileSizeMb
+    if (maxMb > 0 && file.size > maxMb * 1024 * 1024) {
+      error.value = `File too large (max ${maxMb} MB)`
       throw new Error(error.value)
     }
     uploading.value = true

frontend/src/features/document/ui/DocumentUpload.vue

@@ -44,11 +44,14 @@ const fileInput = ref<HTMLInputElement | null>(null)
 const dragging = ref(false)
 
 const uploadHint = computed(() => {
-  const size = t('upload.maxSize')
-  if (flags.maxPageCount > 0) {
-    return `${size} · ${t('upload.maxPages').replace('{n}', String(flags.maxPageCount))}`
+  const parts: string[] = []
+  if (flags.maxFileSizeMb > 0) {
+    parts.push(t('upload.maxSize').replace('{n}', String(flags.maxFileSizeMb)))
   }
-  return size
+  if (flags.maxPageCount > 0) {
+    parts.push(t('upload.maxPages').replace('{n}', String(flags.maxPageCount)))
+  }
+  return parts.join(' · ')
 })
 
 function openFilePicker() {

frontend/src/features/feature-flags/store.test.ts

@@ -68,6 +68,20 @@ describe('useFeatureFlagStore', () => {
     expect(store.isEnabled('disclaimer')).toBe(false)
   })
 
+  it('reads maxFileSizeMb from health response', async () => {
+    mockApiFetch.mockResolvedValue({ status: 'ok', engine: 'local', maxFileSizeMb: 100 })
+    const store = useFeatureFlagStore()
+    await store.load()
+    expect(store.maxFileSizeMb).toBe(100)
+  })
+
+  it('defaults maxFileSizeMb to 0 when missing', async () => {
+    mockApiFetch.mockResolvedValue({ status: 'ok', engine: 'local' })
+    const store = useFeatureFlagStore()
+    await store.load()
+    expect(store.maxFileSizeMb).toBe(0)
+  })
+
   it('handles health endpoint failure gracefully', async () => {
     mockApiFetch.mockRejectedValue(new Error('Network error'))
     const store = useFeatureFlagStore()

frontend/src/features/feature-flags/store.ts

@@ -10,6 +10,7 @@ interface HealthResponse {
   engine: ConversionEngine
   deploymentMode?: DeploymentMode
   maxPageCount?: number
+  maxFileSizeMb?: number
 }
 
 export type FeatureFlag = 'chunking' | 'disclaimer'
@@ -39,6 +40,7 @@ export const useFeatureFlagStore = defineStore('feature-flags', () => {
   const engine = ref<ConversionEngine | null>(null)
   const deploymentMode = ref<DeploymentMode | null>(null)
   const maxPageCount = ref<number>(0)
+  const maxFileSizeMb = ref<number>(0)
   const loaded = ref(false)
   const error = ref<string | null>(null)
 
@@ -59,6 +61,7 @@ export const useFeatureFlagStore = defineStore('feature-flags', () => {
       engine.value = data.engine
       deploymentMode.value = data.deploymentMode ?? 'self-hosted'
       maxPageCount.value = data.maxPageCount ?? 0
+      maxFileSizeMb.value = data.maxFileSizeMb ?? 0
       loaded.value = true
       error.value = null
     } catch (e) {
@@ -67,5 +70,5 @@ export const useFeatureFlagStore = defineStore('feature-flags', () => {
     }
   }
 
-  return { engine, deploymentMode, maxPageCount, loaded, error, isEnabled, load }
+  return { engine, deploymentMode, maxPageCount, maxFileSizeMb, loaded, error, isEnabled, load }
 })

frontend/src/shared/i18n.ts

@@ -94,9 +94,9 @@ const messages: Messages = {
     // Upload
     'upload.drop': 'Déposez un PDF ici ou cliquez pour importer',
     'upload.uploading': 'Import en cours...',
-    'upload.maxSize': 'Max 5Mo',
+    'upload.maxSize': 'Max {n}Mo',
     'upload.invalidFormat': 'Format invalide — seuls les fichiers PDF sont acceptés.',
-    'upload.tooLarge': 'Fichier trop volumineux (max 5 Mo).',
+    'upload.tooLarge': 'Fichier trop volumineux (max {n} Mo).',
     'upload.maxPages': 'Max {n} pages',
 
     // History
@@ -135,7 +135,7 @@ const messages: Messages = {
 
     // Disclaimer
     'disclaimer.banner':
-      'Instance de d\u00e9monstration \u2014 les documents upload\u00e9s sont partag\u00e9s et temporaires (max 5 Mo). Ne pas envoyer de fichiers confidentiels.',
+      'Instance de d\u00e9monstration \u2014 les documents upload\u00e9s sont partag\u00e9s et temporaires (max {n} Mo). Ne pas envoyer de fichiers confidentiels.',
   },
   en: {
     'nav.home': 'Home',
@@ -218,9 +218,9 @@ const messages: Messages = {
 
     'upload.drop': 'Drop a PDF here or click to upload',
     'upload.uploading': 'Uploading...',
-    'upload.maxSize': 'Max 5MB',
+    'upload.maxSize': 'Max {n}MB',
     'upload.invalidFormat': 'Invalid format — only PDF files are accepted.',
-    'upload.tooLarge': 'File too large (max 5 MB).',
+    'upload.tooLarge': 'File too large (max {n} MB).',
     'upload.maxPages': 'Max {n} pages',
 
     'history.title': 'History',
@@ -255,7 +255,7 @@ const messages: Messages = {
 
     // Disclaimer
     'disclaimer.banner':
-      'Demo instance \u2014 uploaded documents are shared and temporary (max 5 MB). Do not upload confidential files.',
+      'Demo instance \u2014 uploaded documents are shared and temporary (max {n} MB). Do not upload confidential files.',
   },
 }