Upload 2 files

app.py

@@ -0,0 +1,50 @@
+import gradio as gr
+
+# --- Minimal Sanskrit lexicon (extend with real data) ---
+LEXICON = {
+    "राम", "वन", "गच्छति", "गुरु", "इन्द्र", "तत्", "अपि",
+    "धर्म", "क्षेत्र", "कुरु", "क्षेत्रे"
+}
+
+# --- Basic Reverse Sandhi Rules ---
+REVERSE_SANDHI_RULES = [
+    ("ा", ["अ+अ"]),    # ā → a + a
+    ("े", ["अ+इ", "अ+ई"]),  # e → a+i or a+ī
+    ("ो", ["अ+उ", "अ+ऊ"]),  # o → a+u or a+ū
+    ("ः", ["ः+"]),          # visarga restoration
+]
+
+def generate_candidates(word):
+    candidates = []
+    for i in range(1, len(word)):
+        left, right = word[:i], word[i:]
+        # Direct split
+        if left in LEXICON and right in LEXICON:
+            candidates.append((left, right))
+        # Apply reverse sandhi substitutions
+        for ch, expansions in REVERSE_SANDHI_RULES:
+            if left.endswith(ch):
+                for exp in expansions:
+                    l_base = left[:-1] + exp.split("+")[0]
+                    r_base = exp.split("+")[1] + right
+                    if l_base in LEXICON and r_base in LEXICON:
+                        candidates.append((l_base, r_base))
+    # Deduplicate
+    candidates = list(set(candidates))
+    return candidates or [("No plausible split found", "")]
+
+def sandhi_splitter(word):
+    candidates = generate_candidates(word.strip())
+    formatted = [" + ".join(c) for c in candidates]
+    return "\n".join(formatted)
+
+with gr.Blocks() as demo:
+    gr.Markdown("## Sanskrit Sandhi-Splitter (Prototype)")
+    gr.Markdown("Enter a Sanskrit compound word (Devanagari) to see possible splits.")
+    inp = gr.Textbox(label="Compound Word (e.g. धर्मक्षेत्रे)")
+    out = gr.Textbox(label="Candidate Splits")
+    btn = gr.Button("Split Sandhi")
+    btn.click(fn=sandhi_splitter, inputs=inp, outputs=out)
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1 @@
+gradio