Split conversation transcript parsing

src/conversation/codex.rs

@@ -0,0 +1,105 @@
+use serde_json::Value;
+
+use super::{ConversationSource, TranscriptMessage, TranscriptRole};
+
+pub struct CodexSource;
+
+impl ConversationSource for CodexSource {
+    fn parse_message(&self, obj: &Value, offset: u64) -> Option<TranscriptMessage> {
+        parse_message(obj, offset)
+    }
+
+    fn is_compaction(&self, obj: &Value) -> bool {
+        is_compaction(obj)
+    }
+
+    fn may_contain_compaction(&self, obj_bytes: &[u8]) -> bool {
+        contains_bytes(obj_bytes, b"context_compacted")
+    }
+}
+
+pub(crate) fn parse_message(obj: &Value, offset: u64) -> Option<TranscriptMessage> {
+    if obj.get("type").and_then(|v| v.as_str()) != Some("event_msg") {
+        return None;
+    }
+
+    let payload = obj.get("payload")?;
+    let (role, text) = match payload.get("type").and_then(|v| v.as_str()) {
+        Some("user_message") => (
+            TranscriptRole::User,
+            payload.get("message").and_then(|v| v.as_str())?.to_string(),
+        ),
+        Some("agent_message") => (
+            TranscriptRole::Assistant,
+            payload.get("message").and_then(|v| v.as_str())?.to_string(),
+        ),
+        _ => return None,
+    };
+
+    if text.is_empty() {
+        return None;
+    }
+
+    let timestamp = obj.get("timestamp")
+        .and_then(|v| v.as_str())
+        .map(str::to_string);
+
+    Some(TranscriptMessage { role, text, timestamp, offset })
+}
+
+pub(crate) fn is_compaction(obj: &Value) -> bool {
+    obj.get("type").and_then(|v| v.as_str()) == Some("event_msg")
+        && obj.get("payload")
+            .and_then(|p| p.get("type"))
+            .and_then(|v| v.as_str()) == Some("context_compacted")
+}
+
+fn contains_bytes(haystack: &[u8], needle: &[u8]) -> bool {
+    haystack.windows(needle.len()).any(|w| w == needle)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use serde_json::json;
+
+    #[test]
+    fn parses_event_messages_and_skips_noise() {
+        let user = json!({
+            "timestamp": "2026-06-15T15:00:00.000Z",
+            "type": "event_msg",
+            "payload": { "type": "user_message", "message": "start here" }
+        });
+        let assistant = json!({
+            "timestamp": "2026-06-15T15:00:01.000Z",
+            "type": "event_msg",
+            "payload": { "type": "agent_message", "message": "working" }
+        });
+        let tool = json!({
+            "timestamp": "2026-06-15T15:00:02.000Z",
+            "type": "event_msg",
+            "payload": { "type": "task_started" }
+        });
+        let raw = json!({
+            "timestamp": "2026-06-15T15:00:03.000Z",
+            "type": "response_item",
+            "payload": { "type": "message", "role": "user" }
+        });
+
+        assert_eq!(parse_message(&user, 1).unwrap().role, TranscriptRole::User);
+        assert_eq!(parse_message(&assistant, 2).unwrap().text, "working");
+        assert!(parse_message(&tool, 3).is_none());
+        assert!(parse_message(&raw, 4).is_none());
+    }
+
+    #[test]
+    fn detects_compaction_event() {
+        let obj = json!({
+            "timestamp": "2026-06-15T15:00:01.000Z",
+            "type": "event_msg",
+            "payload": { "type": "context_compacted" }
+        });
+
+        assert!(is_compaction(&obj));
+    }
+}