feat: :sparkles: Add lang filter

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 
 from back_end import next_annotation, submit_correction, enable_buttons
-from utils import diff_texts
+from utils import diff_texts, LANGS
 
 
 # Gradio Interface
@@ -31,6 +31,13 @@ with gr.Blocks() as demo:
                    label="OFF Password",
                    type="password",
                 )
+                lang = gr.Dropdown(
+                    choices=LANGS.keys(),
+                    value="All",
+                    label="Language", 
+                    info="Optional: Filter a language.",
+                    
+                )
 
             # Saved to detect change from annotator
             model_correction = gr.Text(visible=False)
@@ -54,7 +61,7 @@ with gr.Blocks() as demo:
                 label="Difference Between Original and Corrected Text",
                 combine_adjacent=True,
                 show_legend=True,
-                color_map={"+": "green", "-": "red"}  # "+" for inserted text, "-" for deleted text
+                color_map={"+": "green", "-": "red"},
             )
 
         image = gr.Image()
@@ -66,26 +73,26 @@ with gr.Blocks() as demo:
     
     validate_button.click(
         submit_correction,
-        inputs=[insight_id, annotator_correction, model_correction, off_username, off_password],
-        outputs=[insight_id, original, model_correction, annotator_correction, image]
+        inputs=[insight_id, annotator_correction, model_correction, off_username, off_password, lang],
+        outputs=[insight_id, original, model_correction, annotator_correction, image],
     )
 
     skip_button.click(
         next_annotation,
-        inputs=[],
-        outputs=[insight_id, original, model_correction, annotator_correction, image]
+        inputs=[lang],
+        outputs=[insight_id, original, model_correction, annotator_correction, image],
     )
     
     annotator_correction.change(
         diff_texts,  # Call diff function
         inputs=[original, annotator_correction],
-        outputs=diff_display
+        outputs=diff_display,
     )
 
     off_username.change(
         enable_buttons, 
         inputs=[off_username, off_password], 
-        outputs=[validate_button, skip_button]
+        outputs=[validate_button, skip_button],
     )
     
     off_password.change(
@@ -94,6 +101,12 @@ with gr.Blocks() as demo:
         outputs=[validate_button, skip_button],
     )
 
+    lang.change(
+        next_annotation,
+        inputs=[lang],
+        outputs=[insight_id, original, model_correction, annotator_correction, image],
+    )
+
     # Load the first set of texts when the demo starts
     demo.load(
         next_annotation,

back_end.py

@@ -4,7 +4,7 @@ import base64
 import json
 from dataclasses import dataclass
 
-from utils import get_logger
+from utils import get_logger, LANGS
 
 from openfoodfacts.api import ProductResource, APIConfig
 import gradio as gr
@@ -29,8 +29,11 @@ class Authentification:
         return base64.b64encode(credentials.encode()).decode()
 
 
-def next_annotation() -> Annotation:
-    insight = import_random_insight()
+def next_annotation(lang: Optional[str]) -> Annotation:
+    lang_key = LANGS.get(lang)
+    if not lang_key:
+        logger.warning("The provided Lang was not found. 'All' returned by default.")
+    insight = import_random_insight(lang_key=lang_key)
     logger.info("Imported insight: %s", insight)
     return (
         insight["id"],
@@ -74,13 +77,23 @@ def submit_correction(
 def import_random_insight(
     insight_type: str = "ingredient_spellcheck",
     predictor: str = "fine-tuned-mistral-7b",
+    lang_key: Optional[str] = None
 ) -> Dict:
     url = f"{BASE_URL}/insights/random?count=1&type={insight_type}&predictor={predictor}"
-    response = requests.get(url)
-    data = response.json()
-    insight = data["insights"][0]
-    return insight
-
+    if lang_key:
+        url += f"&value_tag={lang_key}"
+    try:
+        response = requests.get(url)
+        response.raise_for_status()  # Raises an HTTPError for bad responses (4xx or 5xx)
+        data = response.json()
+        if "insights" in data and data["insights"]:
+            return data["insights"][0]
+        else:
+            gr.Warning("No insights found; fetching default insight instead.")
+            return import_random_insight(insight_type, predictor)
+    except requests.RequestException as e:
+        gr.Error(f"Import product from Product Opener failed: {e}")
+        
 
 def submit_to_product_opener(
     insight_id: str,

utils.py

@@ -13,6 +13,65 @@ def diff_texts(text1, text2):
 def get_logger():
     logging.basicConfig(
         level=logging.INFO,
-        format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+        format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
     )
     return logging.getLogger(__name__)
+
+
+LANGS = {
+    "All": None,
+    "English": "en",
+    "French": "fr",
+    "German": "de",
+    "Spanish": "es",
+    "Italian": "it",
+    "Dutch": "nl",
+    "Polish": "pl",
+    "Portuguese": "pt",
+    "Swedish": "sv",
+    "Bulgarian": "bg",
+    "Romanian": "ro",
+    "Finnish": "fi",
+    "Russian": "ru",
+    "Norwegian Bokmål": "nb",
+    "Czech": "cs",
+    "Thai": "th",
+    "Danish": "da",
+    "Croatian": "hr",
+    "Hungarian": "hu",
+    "Arabic": "ar",
+    "Greek": "el",
+    "Japanese": "ja",
+    "Catalan": "ca",
+    "Serbian": "sr",
+    "Slovenian": "sl",
+    "Slovak": "sk",
+    "Turkish": "tr",
+    "Lithuanian": "lt",
+    "Chinese": "zh",
+    "Estonian": "et",
+    "Latvian": "lv",
+    "Undefined": "xx",
+    "Ukrainian": "uk",
+    "Indonesian": "id",
+    "Hebrew": "he",
+    "Vietnamese": "vi",
+    "Icelandic": "is",
+    "Latin": "la",
+    "Korean": "ko",
+    "Albanian": "sq",
+    "Georgian": "ka",
+    "Malay": "ms",
+    "Bosnian": "bs",
+    "Persian": "fa",
+    "Bengali": "bn",
+    "Galician": "gl",
+    "Kazakh": "kk",
+    "Macedonian": "mk",
+    "Norwegian Nynorsk": "nn",
+    "Hindi": "hi",
+    "Afar": "aa",
+    "Uzbek": "uz",
+    "Somali": "so",
+    "Afrikaans": "af"
+}