Merge pull request #68 from x-tabdeveloping/embedding_dim

Informative error message

Author: KennethEnevoldsen

pyproject.toml

@@ -6,7 +6,7 @@ line-length=79
 
 [tool.poetry]
 name = "turftopic"
-version = "0.7.0"
+version = "0.7.1"
 description = "Topic modeling with contextual representations from sentence transformers."
 authors = ["Márton Kardos <power.up1163@gmail.com>"]
 license = "MIT"

turftopic/models/_keynmf.py

@@ -16,6 +16,12 @@
 
 from turftopic.base import Encoder
 
+NOT_MATCHING_ERROR = (
+    "Document embedding dimensionality ({n_dims}) doesn't match term embedding dimensionality ({n_word_dims}). "
+    + "Perhaps you are using precomputed embeddings but forgot to pass an encoder to your model. "
+    + "Try to initialize the model with the encoder you used for computing the embeddings."
+)
+
 
 def batched(iterable, n: int) -> Iterable[list[str]]:
     "Batch data into tuples of length n. The last batch may be shorter."
@@ -143,6 +149,13 @@ def batch_extract_keywords(
                 self.term_embeddings[self.key_to_index[term]]
                 for term in batch_vocab[important_terms]
             ]
+            if self.term_embeddings.shape[1] != embeddings.shape[1]:
+                raise ValueError(
+                    NOT_MATCHING_ERROR.format(
+                        n_dims=embeddings.shape[1],
+                        n_word_dims=self.term_embeddings.shape[1],
+                    )
+                )
             sim = cosine_similarity(embedding, word_embeddings).astype(
                 np.float64
             )

turftopic/models/cluster.py

@@ -39,6 +39,12 @@
 feature_importance must be one of 'soft-c-tf-idf', 'c-tf-idf', 'centroid'
 """
 
+NOT_MATCHING_ERROR = (
+    "Document embedding dimensionality ({n_dims}) doesn't match term embedding dimensionality ({n_word_dims}). "
+    + "Perhaps you are using precomputed embeddings but forgot to pass an encoder to your model. "
+    + "Try to initialize the model with the encoder you used for computing the embeddings."
+)
+
 
 def smallest_hierarchical_join(
     topic_vectors: np.ndarray,
@@ -370,6 +376,16 @@ def estimate_components(
                 self.vocab_embeddings = self.encoder_.encode(
                     self.vectorizer.get_feature_names_out()
                 )  # type: ignore
+                if (
+                    self.vocab_embeddings.shape[1]
+                    != self.topic_vectors_.shape[1]
+                ):
+                    raise ValueError(
+                        NOT_MATCHING_ERROR.format(
+                            n_dims=self.topic_vectors_.shape[1],
+                            n_word_dims=self.vocab_embeddings.shape[1],
+                        )
+                    )
             self.components_ = cluster_centroid_distance(
                 self.topic_vectors_,
                 self.vocab_embeddings,

turftopic/models/decomp.py

@@ -11,6 +11,12 @@
 from turftopic.base import ContextualModel, Encoder
 from turftopic.vectorizer import default_vectorizer
 
+NOT_MATCHING_ERROR = (
+    "Document embedding dimensionality ({n_dims}) doesn't match term embedding dimensionality ({n_word_dims}). "
+    + "Perhaps you are using precomputed embeddings but forgot to pass an encoder to your model. "
+    + "Try to initialize the model with the encoder you used for computing the embeddings."
+)
+
 
 class SemanticSignalSeparation(ContextualModel):
     """Separates the embedding matrix into 'semantic signals' with
@@ -115,6 +121,13 @@ def fit_transform(
             console.log("Term extraction done.")
             status.update("Encoding vocabulary")
             self.vocab_embeddings = self.encoder_.encode(vocab)
+            if self.vocab_embeddings.shape[1] != self.embeddings.shape[1]:
+                raise ValueError(
+                    NOT_MATCHING_ERROR.format(
+                        n_dims=self.embeddings.shape[1],
+                        n_word_dims=self.vocab_embeddings.shape[1],
+                    )
+                )
             console.log("Vocabulary encoded.")
             status.update("Estimating term importances")
             vocab_topic = self.decomposition.transform(self.vocab_embeddings)