Train on everything if train_on is unspecified; add line in docs

sillsdev · Nov 8, 2023 · ffa3206 · ffa3206
1 parent d532be1
commit ffa3206
Show file tree

Hide file tree

Showing 3 changed files with 10 additions and 0 deletions.
diff --git a/src/Serval.Client/Client.g.cs b/src/Serval.Client/Client.g.cs
@@ -1013,6 +1013,8 @@ public partial interface ITranslationEnginesClient
         /// <br/>you may flag a subset of books for pretranslation by including their [abbreviations](https://github.com/sillsdev/libpalaso/blob/master/SIL.Scripture/Canon.cs)
         /// <br/>in the textIds parameter. If the engine does not support pretranslation, these fields have no effect.
         /// <br/>            
+        /// <br/>Similarly, specify the corpora and textIds to train on. If none are provided, all corpora will be used.
+        /// <br/>            
         /// <br/>The `"options"` parameter of the build config provides the ability to pass build configuration parameters as a JSON object.
         /// <br/>A typical use case would be to set `"options"` to `{"max_steps":10}` in order to configure the maximum
         /// <br/>number of training iterations in order to reduce turnaround time for testing purposes.
@@ -2914,6 +2916,8 @@ public string BaseUrl
         /// <br/>you may flag a subset of books for pretranslation by including their [abbreviations](https://github.com/sillsdev/libpalaso/blob/master/SIL.Scripture/Canon.cs)
         /// <br/>in the textIds parameter. If the engine does not support pretranslation, these fields have no effect.
         /// <br/>            
+        /// <br/>Similarly, specify the corpora and textIds to train on. If none are provided, all corpora will be used.
+        /// <br/>            
         /// <br/>The `"options"` parameter of the build config provides the ability to pass build configuration parameters as a JSON object.
         /// <br/>A typical use case would be to set `"options"` to `{"max_steps":10}` in order to configure the maximum
         /// <br/>number of training iterations in order to reduce turnaround time for testing purposes.

diff --git a/src/Serval.Translation/Controllers/TranslationEnginesController.cs b/src/Serval.Translation/Controllers/TranslationEnginesController.cs
@@ -772,6 +772,8 @@ CancellationToken cancellationToken
     /// you may flag a subset of books for pretranslation by including their [abbreviations](https://github.com/sillsdev/libpalaso/blob/master/SIL.Scripture/Canon.cs)
     /// in the textIds parameter. If the engine does not support pretranslation, these fields have no effect.
     ///
+    /// Similarly, specify the corpora and textIds to train on. If none are provided, all corpora will be used.
+    ///
     /// The `"options"` parameter of the build config provides the ability to pass build configuration parameters as a JSON object.
     /// A typical use case would be to set `"options"` to `{"max_steps":10}` in order to configure the maximum
     /// number of training iterations in order to reduce turnaround time for testing purposes.

diff --git a/src/Serval.Translation/Services/EngineService.cs b/src/Serval.Translation/Services/EngineService.cs
@@ -217,6 +217,10 @@ public async Task<bool> StartBuildAsync(Build build, CancellationToken cancellat
                             if (trainingCorpus.TextIds is not null)
                                 corpus.TrainOnTextIds.Add(trainingCorpus.TextIds);
                         }
+                        else if (trainOn?.Count == 0)
+                        {
+                            corpus.TrainOnAll = true;
+                        }
                         return corpus;
                     })
                 }