fix: dedupl user prompt on history

.github/workflows/ci.yml

@@ -1,6 +1,9 @@
 name: CI
 
-on: [push, pull_request]
+on:
+  push:
+    branches: [main]
+  pull_request:
 
 permissions:
   contents: read

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "openai-api-dispatch"
-version = "0.1.0"
+version = "0.1.1"
 authors = ["Victor Lopez <vhrlopes@gmail.com>"]
 description = "OpenAI-compatible chat requests through memory or NATS queues."
 edition = "2024"

README.md

@@ -59,21 +59,32 @@ Environment settings are read when constructing producers, workers, and executor
 | Variable | Default |
 | --- | --- |
 | `OPENAI_API_NATS_URL` | `nats://localhost:4222` |
+| `OPENAI_API_NATS_WORKERS_GROUP` | `task_workers` |
 | `OPENAI_API_NATS_PREFIX` | `openai-api-queue/` |
 | `OPENAI_API_URL` | `http://127.0.0.1:8000/v1` |
 | `OPENAI_API_DEFAULT_MODEL` | Unset; required by NATS workers |
 
-- A task's explicit model overrides the default and selects its NATS subject. Prefix and model are concatenated verbatim. Workers share the `task_workers` queue group; this uses Core NATS without persistence or task retries. Constructors do not wait for server confirmation of subscriptions, so startup can race with publishing.
-- Each worker handles one task at a time. Queue/API errors stop its loop without an error reply; supervise spawned workers. `send_and_wait` limits only the reply wait, not submission, and expiry does not cancel execution. Check `response.success` even when the call returns `Ok`.
-- Only non-streaming chat is implemented. The current prompt is sent as a user message; system entries in input history are ignored (use `with_system`). Returned history currently duplicates the latest user prompt before the assistant reply. Schemas request strict JSON output without local validation. Keep `max_tokens` within `u32`; it is cast unchecked. `payload` is caller metadata, not model input.
+- A task's explicit model overrides the default and selects its NATS subject.
+- Constructors do not wait for server confirmation of subscriptions, so startup can race with publishing.
+- Each worker handles one task at a time.
+- Queue/API errors stop its loop without an error reply; supervise spawned workers.
+- `send_and_wait` limits only the reply wait, not submission, and expiry does not cancel execution.
+- Check `response.success` even when the call returns `Ok`.
+- Only non-streaming chat is implemented.
+- The current prompt is sent as a user message; system entries in input history are ignored (use `with_system`).
+- Schemas request strict JSON output without local validation.
+- `payload` is caller metadata, not model input.
+- `no_std` generated IDs can repeat after restart or wraparound.
 
 ## Features
 
 Default features are `std`, `memory-queue`, and `nats-queue`; NATS implies `std`. The API executor requires `std`.
 
-For `no_std` memory queues, set `default-features = false, features = ["memory-queue"]`. An allocator and pointer/32-bit atomics are required. This backend polls once, has no timeout support, and evicts old items when bounded; the `std` backend uses bounded Tokio channels and backpressure. Use positive capacities and unique task IDs: `TaskBuilder::default()` uses ID zero, and `no_std` generated IDs can repeat after restart or wraparound.
+#### no_std
 
-Custom queues/executors implement `QueueProducer`, `QueueWorker`, and `Executor`. Only the NATS/API pairing has a public `Worker` constructor; other combinations need a caller-managed loop.
+For memory queues, use the feature `memory-queue`. An allocator and pointer/32-bit atomics are required.
+
+This backend polls once, has no timeout support, and evicts old items when bounded; the `std` backend uses bounded Tokio channels and backpressure. 
 
 ## Development
 

src/executor/openai/mod.rs

@@ -195,7 +195,6 @@ impl ExecutorAsyncOpenai {
                                 None
                             }
                         })
-                        .chain(iter::once(Interaction::User(prompt)))
                         .chain(iter::once(Interaction::Assistant(m.clone())))
                         .collect();
 

src/queue/nats/mod.rs

@@ -33,7 +33,7 @@ pub struct NatsProducer {
 }
 
 #[derive(Debug, Clone)]
-/// Consumes one model subject in the fixed `task_workers` queue group.
+/// Consumes one model subject in the workers queue group.
 ///
 /// Clones share one subscription. Replies require the incoming message's reply
 /// subject; malformed JSON and missing reply subjects return errors.
@@ -103,8 +103,10 @@ impl NatsWorker {
             .connect(url)
             .await?;
 
+        let workers_group =
+            env::var("OPENAI_API_NATS_WORKERS_GROUP").unwrap_or_else(|_| "task_workers".into());
         let subscriber = client
-            .queue_subscribe(subject.clone(), "task_workers".to_string())
+            .queue_subscribe(subject.clone(), workers_group)
             .await?;
         let subscriber = Arc::new(Mutex::new(subscriber));
 

src/task.rs

@@ -7,7 +7,7 @@ use serde_json::Value;
 
 use crate::{queue::QueueProducer, utils};
 
-#[derive(Debug, Default, Clone, PartialEq, Eq, Serialize, Deserialize)]
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
 /// Builds a chat task; only the prompt is required at build time.
 ///
 /// Prefer [`Self::new`] for a generated ID: derived `Default` uses ID zero.
@@ -30,12 +30,24 @@ pub struct TaskBuilder {
     pub prompt: Option<String>,
 }
 
+impl Default for TaskBuilder {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
 impl TaskBuilder {
     /// Creates an empty builder with an ID from [`utils::id`].
     pub fn new() -> Self {
         Self {
             id: utils::id(),
-            ..Default::default()
+            model: None,
+            max_tokens: None,
+            payload: None,
+            system: None,
+            history: None,
+            schema: None,
+            prompt: None,
         }
     }