cuekit and cuenorm: three subtitle formats, one cue stream

crates/cuekit/src/cue.rs

@@ -0,0 +1,82 @@
+//! One line of a transcript, and when it was said.
+
+use schemars::JsonSchema;
+use serde::{Deserialize, Serialize};
+
+/// A subtitle cue: a span of time and what was said in it.
+///
+/// Deliberately smaller than a chunk. A cue is what a subtitle file holds —
+/// a few seconds, often half a sentence — and turning cues into passages
+/// worth retrieving is a later stage's job.
+#[derive(Debug, Clone, PartialEq, Serialize, Deserialize, JsonSchema)]
+pub struct Cue {
+    /// Record version.
+    #[serde(default = "version")]
+    pub v: u32,
+    /// Seconds from the start of the recording.
+    #[serde(with = "corpus_core::finite")]
+    #[schemars(with = "f64")]
+    pub start: f64,
+    /// Seconds from the start of the recording.
+    #[serde(with = "corpus_core::finite")]
+    #[schemars(with = "f64")]
+    pub end: f64,
+    /// What was said, with the markup taken out.
+    pub text: String,
+}
+
+fn version() -> u32 {
+    corpus_core::V
+}
+
+impl Cue {
+    /// A cue.
+    pub fn new(start: f64, end: f64, text: impl Into<String>) -> Self {
+        Self {
+            v: corpus_core::V,
+            start,
+            end,
+            text: text.into(),
+        }
+    }
+
+    /// How long the cue lasts. Negative if the file says it ends before it
+    /// starts, which some do.
+    pub fn seconds(&self) -> f64 {
+        self.end - self.start
+    }
+}
+
+impl corpus_core::Record for Cue {
+    fn version(&self) -> u32 {
+        self.v
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn every_field_is_emitted() {
+        let cue = Cue::new(12.5, 15.0, "the mill stood at the northern end");
+        assert_eq!(
+            serde_json::to_string(&cue).unwrap(),
+            r#"{"v":1,"start":12.5,"end":15.0,"text":"the mill stood at the northern end"}"#
+        );
+    }
+
+    #[test]
+    fn a_cue_reads_back_identical() {
+        let cue = Cue::new(0.0, 1.0 / 3.0, "x");
+        let text = serde_json::to_string(&cue).unwrap();
+        let back: Cue = serde_json::from_str(&text).unwrap();
+        assert_eq!(back.end.to_bits(), cue.end.to_bits());
+    }
+
+    #[test]
+    fn a_time_that_json_cannot_carry_is_refused() {
+        let cue = Cue::new(f64::NAN, 1.0, "x");
+        assert!(serde_json::to_string(&cue).is_err());
+    }
+}