Core Prakriya Engine and Pilot Derivation Rules

rust/vedyut-cheda/Cargo.toml

@@ -7,9 +7,9 @@ repository.workspace = true
 description = "Sanskrit text segmentation and morphological analysis"
 
 [dependencies]
-vedyut-lipi = { path = "../vedyut-lipi" }
-vedyut-sandhi = { path = "../vedyut-sandhi" }
-vedyut-kosha = { path = "../vedyut-kosha" }
+vedyut-lipi = { path = "../vedyut-lipi", version = "0.1.0" }
+vedyut-sandhi = { path = "../vedyut-sandhi", version = "0.1.0" }
+vedyut-kosha = { path = "../vedyut-kosha", version = "0.1.0" }
 serde = { workspace = true }
 rustc-hash = { workspace = true }
 

rust/vedyut-core/Cargo.toml

@@ -11,12 +11,12 @@ name = "vedyut_core"
 crate-type = ["cdylib", "rlib"]
 
 [dependencies]
-vedyut-lipi = { path = "../vedyut-lipi" }
-vedyut-sandhi = { path = "../vedyut-sandhi" }
-vedyut-prakriya = { path = "../vedyut-prakriya" }
-vedyut-kosha = { path = "../vedyut-kosha" }
-vedyut-cheda = { path = "../vedyut-cheda" }
-vedyut-sanskritify = { path = "../vedyut-sanskritify" }
+vedyut-lipi = { path = "../vedyut-lipi", version = "0.1.0" }
+vedyut-sandhi = { path = "../vedyut-sandhi", version = "0.1.0" }
+vedyut-prakriya = { path = "../vedyut-prakriya", version = "0.1.0" }
+vedyut-kosha = { path = "../vedyut-kosha", version = "0.1.0" }
+vedyut-cheda = { path = "../vedyut-cheda", version = "0.1.0" }
+vedyut-sanskritify = { path = "../vedyut-sanskritify", version = "0.1.0" }
 
 # PyO3 for Python bindings
 pyo3 = { workspace = true }

rust/vedyut-core/src/lib.rs

@@ -8,7 +8,7 @@ use vedyut_lipi::Scheme;
 
 /// Python module for vedyut
 #[pymodule]
-fn _core(_py: Python, m: &PyModule) -> PyResult<()> {
+fn _core(m: &Bound<'_, PyModule>) -> PyResult<()> {
     // Register classes and functions
     m.add_class::<PyScheme>()?;
     m.add_function(wrap_pyfunction!(py_transliterate, m)?)?;
@@ -136,7 +136,7 @@ fn py_analyze(word: &str, script: &str, py: Python) -> PyResult<Vec<PyObject>> {
     })?;
 
     if let Some(analysis) = vedyut_cheda::analyze_word(word) {
-        let dict = PyDict::new(py);
+        let dict = PyDict::new_bound(py);
         dict.set_item("word", analysis.word)?;
         dict.set_item("stem", analysis.stem)?;
         dict.set_item("linga", analysis.linga)?;
@@ -158,8 +158,8 @@ mod tests {
     fn test_module_creation() {
         pyo3::prepare_freethreaded_python();
         Python::with_gil(|py| {
-            let module = PyModule::new(py, "_core").unwrap();
-            assert!(_core(py, module).is_ok());
+            let module = PyModule::new_bound(py, "_core").unwrap();
+            assert!(_core(&module).is_ok());
         });
     }
 }

rust/vedyut-kosha/Cargo.toml

@@ -7,7 +7,7 @@ repository.workspace = true
 description = "High-performance lexicon for Sanskrit"
 
 [dependencies]
-vedyut-lipi = { path = "../vedyut-lipi" }
+vedyut-lipi = { path = "../vedyut-lipi", version = "0.1.0" }
 serde = { workspace = true }
 serde_json = { workspace = true }
 rustc-hash = { workspace = true }

rust/vedyut-prakriya/Cargo.toml

@@ -7,7 +7,7 @@ repository.workspace = true
 description = "Paninian word generator for Sanskrit"
 
 [dependencies]
-vedyut-lipi = { path = "../vedyut-lipi" }
+vedyut-lipi = { path = "../vedyut-lipi", version = "0.1.0" }
 serde = { workspace = true }
 serde_json = { workspace = true }
 compact_str = { workspace = true }

rust/vedyut-prakriya/src/ac_sandhi.rs

@@ -0,0 +1,226 @@
+use crate::prakriya::Prakriya;
+use crate::tag::Tag;
+
+/// 6.1.101 akaḥ savarṇe dīrghaḥ
+pub fn rule_6_1_101(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let mut i = 0;
+    while i < p.terms.len().saturating_sub(1) {
+        let left = &p.terms[i];
+        let right = &p.terms[i + 1];
+
+        // Simplified check, normally would check Ac
+        let result = match (left.text.as_str(), right.text.as_str()) {
+            ("a", "a") | ("a", "A") | ("A", "a") | ("A", "A") => Some("A"),
+            ("i", "i") | ("i", "I") | ("I", "i") | ("I", "I") => Some("I"),
+            ("u", "u") | ("u", "U") | ("U", "u") | ("U", "U") => Some("U"),
+            _ => None,
+        };
+
+        if let Some(res) = result {
+             p.terms[i].text = res.to_string();
+             p.terms[i].add_tag(Tag::Guna); // Technically Dirgha
+             p.terms.remove(i + 1);
+             p.add_rule("6.1.101 akaḥ savarṇe dīrghaḥ");
+             changed = true;
+        } else {
+            i += 1;
+        }
+    }
+    changed
+}
+
+/// 6.1.87 ādguṇaḥ
+pub fn rule_6_1_87(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let mut i = 0;
+    while i < p.terms.len().saturating_sub(1) {
+        let left_text = p.terms[i].text.clone();
+        let right_text = p.terms[i+1].text.clone();
+
+        let left_last = left_text.chars().last().unwrap();
+        let right_first = right_text.chars().next().unwrap();
+
+        // a/A + i/u/r/l -> e/o/ar/al
+        // Condition: following is ac
+        if "aA".contains(left_last) && "iIuUfFxX".contains(right_first) {
+             let replacement = match right_first {
+                 'i' | 'I' => "e",
+                 'u' | 'U' => "o",
+                 'f' | 'F' => "ar",
+                 'x' | 'X' => "al",
+                 _ => "",
+             };
+
+             if !replacement.is_empty() {
+                 let new_left = format!("{}{}", &left_text[..left_text.len()-left_last.len_utf8()], replacement);
+                 // Note: right term needs to lose its first char?
+                 // Sandhi merges two sounds.
+                 // Left loses last, Right loses first, both replaced by replacement.
+                 // My previous implementations were a bit simplified.
+                 // Correct logic: Merge (last + first) -> replacement
+
+                 p.terms[i].text = new_left;
+                 // Need to remove first char of right term, or merge entirely if it's single char
+                 // Here we simplify by merging entire terms if right is single char, or handling string manipulation
+                 // For subanta/tinanta, right terms are often suffixes starting with vowel
+
+                 // Let's assume right term loses its first char
+                 let new_right = right_text[right_first.len_utf8()..].to_string();
+                 if new_right.is_empty() {
+                     p.terms.remove(i+1);
+                 } else {
+                     p.terms[i+1].text = new_right;
+                 }
+
+                 p.add_rule("6.1.87 ādguṇaḥ");
+                 changed = true;
+             }
+        }
+        i += 1;
+    }
+    changed
+}
+
+/// 6.1.88 vṛddhireci
+pub fn rule_6_1_88(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let mut i = 0;
+    while i < p.terms.len().saturating_sub(1) {
+        let left_text = p.terms[i].text.clone();
+        let right_text = p.terms[i+1].text.clone();
+
+        let left_last = left_text.chars().last().unwrap();
+        let right_first = right_text.chars().next().unwrap();
+
+        // a/A + e/o/ai/au -> ai/au
+        if "aA".contains(left_last) && "eEoO".contains(right_first) {
+             let replacement = match right_first {
+                 'e' | 'E' => "E", // ai
+                 'o' | 'O' => "O", // au
+                 _ => "",
+             };
+
+             if !replacement.is_empty() {
+                 let new_left = format!("{}{}", &left_text[..left_text.len()-left_last.len_utf8()], replacement);
+
+                 p.terms[i].text = new_left;
+                 let new_right = right_text[right_first.len_utf8()..].to_string();
+                 if new_right.is_empty() {
+                     p.terms.remove(i+1);
+                 } else {
+                     p.terms[i+1].text = new_right;
+                 }
+
+                 p.add_rule("6.1.88 vṛddhireci");
+                 changed = true;
+             }
+        }
+        i += 1;
+    }
+    changed
+}
+
+
+/// 6.1.77 iko yaṇaci
+pub fn rule_6_1_77(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let mut i = 0;
+    while i < p.terms.len().saturating_sub(1) {
+        let left_text = p.terms[i].text.clone();
+        let right_text = p.terms[i+1].text.clone();
+
+        let left_last = left_text.chars().last().unwrap();
+        let right_first = right_text.chars().next().unwrap();
+
+        if "iIuUfFxX".contains(left_last) && "aAiIuUfFxXeEoO".contains(right_first) {
+             let replacement = match left_last {
+                 'i' | 'I' => "y",
+                 'u' | 'U' => "v",
+                 'f' | 'F' => "r",
+                 'x' | 'X' => "l",
+                 _ => "",
+             };
+
+             if !replacement.is_empty() {
+                 let new_left = format!("{}{}", &left_text[..left_text.len()-left_last.len_utf8()], replacement);
+                 p.terms[i].text = new_left;
+                 p.add_rule("6.1.77 iko yaṇaci");
+                 changed = true;
+             }
+        }
+        i += 1;
+    }
+    changed
+}
+
+/// 8.2.66 sasajuṣo ruḥ
+pub fn rule_8_2_66(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    // Iterate indices to avoid borrowing conflict
+    for i in 0..p.terms.len() {
+        let term = &p.terms[i];
+        if term.text.ends_with('s') && term.has_tag(Tag::Pada) {
+             p.terms[i].text.pop();
+             p.terms[i].text.push_str("ru~");
+             p.add_rule("8.2.66 sasajuṣo ruḥ");
+             changed = true;
+        }
+    }
+    changed
+}
+
+/// 6.1.78 eco'yavāyāvaḥ
+pub fn rule_6_1_78(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let mut i = 0;
+    while i < p.terms.len().saturating_sub(1) {
+        let left_text = p.terms[i].text.clone();
+        let right_text = p.terms[i+1].text.clone();
+
+        // Simplified check: if left ends in e/o/ai/au and right starts with vowel
+        let last = left_text.chars().last().unwrap();
+        let first = right_text.chars().next().unwrap();
+
+        if "eEoO".contains(last) && "aAiIuUfFxXeEoO".contains(first) {
+            let replacement = match last {
+                'e' => "ay",
+                'o' => "av",
+                'E' => "Ay",
+                'O' => "Av",
+                _ => "",
+            };
+
+            if !replacement.is_empty() {
+                let new_left = format!("{}{}", &left_text[..left_text.len()-last.len_utf8()], replacement);
+                p.terms[i].text = new_left;
+                p.add_rule("6.1.78 eco'yavāyāvaḥ");
+                changed = true;
+            }
+        }
+        i += 1;
+    }
+    changed
+}
+
+/// 8.3.15 kharavasānayorvisarjanīyaḥ
+pub fn rule_8_3_15(p: &mut Prakriya) -> bool {
+    let mut changed = false;
+    let len = p.terms.len();
+    if len > 0 {
+        let last_idx = len - 1;
+        let text = &p.terms[last_idx].text;
+        if text.ends_with("r") || text.ends_with("ru~") {
+             if text.ends_with("ru~") {
+                 let new_text = text.trim_end_matches("ru~").to_string();
+                 p.terms[last_idx].text = new_text;
+             } else {
+                 p.terms[last_idx].text.pop();
+             }
+             p.terms[last_idx].text.push('H');
+             p.add_rule("8.3.15 kharavasānayorvisarjanīyaḥ");
+             changed = true;
+        }
+    }
+    changed
+}