Create app.py

app.py

@@ -0,0 +1,78 @@
+# app.py — CSS LexiScan MVP (Hugging Face Space)
+
+import gradio as gr
+import spacy
+from fuzzywuzzy import fuzz
+from easyocr import Reader
+
+# Load SpaCy small English model
+nlp = spacy.load("en_core_web_sm")
+
+# Initialize EasyOCR reader for English
+reader = Reader(lang_list=['en'])
+
+# Seed idioms dataset
+idioms_dataset = [
+    "break the ice", "spill the beans", "once in a blue moon",
+    "at the drop of a hat", "bark up the wrong tree",
+    "come up with", "look forward to"
+]
+
+# Function to extract difficult vocabulary
+def extract_vocab(text, n=7):
+    doc = nlp(text)
+    candidates = [token.text for token in doc if token.is_alpha and not token.ent_type_ and len(token.text) > 5]
+    unique_words = list(dict.fromkeys(candidates))
+    return unique_words[:n]
+
+# Function to detect idioms
+def detect_idioms(text):
+    found = []
+    doc = nlp(text)
+    sentences = [sent.text.strip() for sent in doc.sents]
+    for idiom in idioms_dataset:
+        for sent in sentences:
+            if fuzz.partial_ratio(idiom.lower(), sent.lower()) > 80:
+                found.append(idiom)
+    return list(set(found))
+
+# Simple CSS-style précis (first 3 sentences)
+def generate_precis(text):
+    doc = nlp(text)
+    sentences = [sent.text.strip() for sent in doc.sents]
+    return " ".join(sentences[:3])
+
+# Main function to process uploaded image
+def process_input(image):
+    text = " ".join(reader.readtext(image, detail=0))
+    precis = generate_precis(text)
+    vocab_list = extract_vocab(text)
+    idioms_found = detect_idioms(text)
+    return (
+        precis,
+        "\n".join(f"- {w}" for w in vocab_list),
+        "\n".join(f"- {i}" for i in idioms_found)
+    )
+
+# Gradio UI
+with gr.Blocks() as demo:
+    gr.Markdown("# 📰 CSS LexiScan — AI-Powered Editorial Analysis")
+    gr.Markdown("Upload a newspaper screenshot to get CSS-style précis, difficult vocabulary, and idioms detected.")
+
+    with gr.Row():
+        with gr.Column():
+            gr.Markdown("### ✍️ CSS-style Précis")
+            precis_box = gr.Textbox(lines=12, placeholder="Generated CSS-style précis will appear here...")
+        with gr.Column():
+            gr.Markdown("### 📚 Difficult Vocabulary")
+            vocab_box = gr.Textbox(lines=8, placeholder="Difficult words will appear here...")
+        with gr.Column():
+            gr.Markdown("### 💡 Idioms Detected")
+            idioms_box = gr.Textbox(lines=8, placeholder="Idioms will appear here...")
+
+    img_input = gr.Image(type="filepath")
+    submit_btn = gr.Button("Analyze Article 📝")
+
+    submit_btn.click(process_input, inputs=img_input, outputs=[precis_box, vocab_box, idioms_box])
+
+demo.launch()