added tests of remaining transformers

packages/langchain-graph-retriever/pyproject.toml

@@ -82,6 +82,7 @@ simsimd = "simsimd"
 spacy = "spacy"
 testcontainers = "testcontainers"
 tqdm = "tqdm"
+types-beautifulsoup4 = "types_beautifulsoup4"
 typing-extensions = "typing_extensions"
 
 [tool.deptry.per_rule_ignores]
@@ -96,7 +97,7 @@ astra = [
     "httpx>=0.28.1",
     "langchain-astradb>=0.5.3",
 ]
-beautifulsoup4 = [
+html = [
     "beautifulsoup4>=4.12.3",
 ]
 cassandra = [
@@ -139,5 +140,6 @@ dev = [
     "simsimd>=6.2.1",
     "testcontainers>=4.9.0",
     "langchain-tests>=0.3.8",
+    "types-beautifulsoup4>=4.12.0.20250204",
 ]
 

packages/langchain-graph-retriever/src/langchain_graph_retriever/adapters/langchain.py

@@ -15,7 +15,7 @@
 )
 
 from langchain_graph_retriever._conversion import doc_to_content
-from langchain_graph_retriever.document_transformers.metadata_denormalizer import (
+from langchain_graph_retriever.transformers.metadata_denormalizer import (
     MetadataDenormalizer,
 )
 

packages/langchain-graph-retriever/src/langchain_graph_retriever/document_transformers/gliner.py → packages/langchain-graph-retriever/src/langchain_graph_retriever/transformers/gliner.py

@@ -1,15 +1,16 @@
 from collections.abc import Sequence
 from typing import Any
 
+from gliner import GLiNER  # type: ignore
 from langchain_core.documents import BaseDocumentTransformer, Document
 from typing_extensions import override
 
 
-class GLiNEREntityExtractor(BaseDocumentTransformer):
+class GLiNERTransformer(BaseDocumentTransformer):
     """
-    Add metadata to documents about named entities using `GLiNER`_.
+    Add metadata to documents about named entities using `GLiNER`.
 
-    `GLiNER`_ is a Named Entity Recognition (NER) model capable of identifying any
+    `GLiNER` is a Named Entity Recognition (NER) model capable of identifying any
     entity type using a bidirectional transformer encoder (BERT-like).
 
     Preliminaries
@@ -20,9 +21,9 @@ class GLiNEREntityExtractor(BaseDocumentTransformer):
     Note that ``bs4`` is also installed to support the WebBaseLoader in the example,
     but not needed by the GLiNEREntityExtractor itself.
 
-    .. code-block:: bash
-
-        pip install -q langchain_community bs4 gliner
+    ```
+    pip install -q langchain_community bs4 gliner
+    ```
 
     Example
     -------
@@ -62,8 +63,9 @@ class GLiNEREntityExtractor(BaseDocumentTransformer):
         A prefix to add to metadata keys outputted by the extractor.
         This will be prepended to the label, with the value (or values) holding the
         generated keywords for that entity kind.
-    model : str, default "urchade/gliner_mediumv2.1"
-        The GLiNER model to use.
+    model : str | GLiNER, default "urchade/gliner_mediumv2.1"
+        The GLiNER model to use. Pass the name of the model to load
+        or pass an instantiated GLiNER model instance.
 
     """  # noqa: E501
 
@@ -73,18 +75,14 @@ def __init__(
         *,
         batch_size: int = 8,
         metadata_key_prefix: str = "",
-        model: str = "urchade/gliner_mediumv2.1",
+        model: Any = "urchade/gliner_mediumv2.1",
     ):
-        try:
-            from gliner import GLiNER  # type: ignore
-
+        if isinstance(model, GLiNER):
+            self._model = model
+        elif isinstance(model, str):
             self._model = GLiNER.from_pretrained(model)
-
-        except ImportError:
-            raise ImportError(
-                "gliner is required for the GLiNEREntityExtractor. "
-                "Please install it with `pip install gliner`."
-            ) from None
+        else:
+            raise ValueError(f"Invalid model: {model}")
 
         self._batch_size = batch_size
         self._labels = labels

packages/langchain-graph-retriever/src/langchain_graph_retriever/document_transformers/html_hyperlink.py → packages/langchain-graph-retriever/src/langchain_graph_retriever/transformers/html_hyperlink.py

@@ -1,20 +1,16 @@
 from __future__ import annotations
 
 from collections.abc import Sequence
-from typing import TYPE_CHECKING, Any
+from typing import Any
 from urllib.parse import urldefrag, urljoin, urlparse
 
-from langchain_core._api import beta
+from bs4 import BeautifulSoup
+from bs4.element import Tag
 from langchain_core.documents import BaseDocumentTransformer, Document
 from typing_extensions import override
 
-if TYPE_CHECKING:
-    from bs4 import BeautifulSoup  # type: ignore
-    from bs4.element import Tag  # type: ignore
 
-
-@beta()
-class HtmlHyperlinkExtractor(BaseDocumentTransformer):
+class HtmlHyperlinkTransformer(BaseDocumentTransformer):
     """
     Extract hyperlinks from HTML content.
 
@@ -207,14 +203,6 @@ def __init__(
         metadata_key: str = "hyperlink",
         drop_fragments: bool = True,
     ):
-        try:
-            from bs4 import BeautifulSoup  # noqa:F401
-        except ImportError as e:
-            raise ImportError(
-                "BeautifulSoup4 is required for HtmlHyperlinkExtractor. "
-                "Please install it with `pip install beautifulsoup4`."
-            ) from e
-
         self._url_metadata_key = url_metadata_key
         self._metadata_key = metadata_key
         self._drop_fragments = drop_fragments
@@ -224,20 +212,25 @@ def _parse_url(link: Tag, page_url: str, drop_fragments: bool = True) -> str | N
         href = link.get("href")
         if href is None:
             return None
+        if isinstance(href, list) and len(href) == 1:
+            href = href[0]
+        if not isinstance(href, str):
+            return None
+
         url = urlparse(href)
         if url.scheme not in ["http", "https", ""]:
             return None
 
         # Join the HREF with the page_url to convert relative paths to absolute.
-        url = str(urljoin(page_url, href))
+        joined_url = str(urljoin(page_url, href))
 
         # Fragments would be useful if we chunked a page based on section.
         # Then, each chunk would have a different URL based on the fragment.
         # Since we aren't doing that yet, they just "break" links. So, drop
         # the fragment.
         if drop_fragments:
-            return urldefrag(url).url
-        return url
+            return urldefrag(joined_url).url
+        return joined_url
 
     @staticmethod
     def _parse_urls(
@@ -247,7 +240,7 @@ def _parse_urls(
         urls: set[str] = set()
 
         for link in soup_links:
-            parsed_url = HtmlHyperlinkExtractor._parse_url(
+            parsed_url = HtmlHyperlinkTransformer._parse_url(
                 link, page_url=page_url, drop_fragments=drop_fragments
             )
             # Remove self links and entries for any 'a' tag that failed to parse

packages/langchain-graph-retriever/src/langchain_graph_retriever/document_transformers/keybert.py → packages/langchain-graph-retriever/src/langchain_graph_retriever/transformers/keybert.py

@@ -1,11 +1,12 @@
 from collections.abc import Sequence
 from typing import Any
 
+from keybert import KeyBERT  # type: ignore
 from langchain_core.documents import BaseDocumentTransformer, Document
 from typing_extensions import override
 
 
-class KeybertKeywordExtractor(BaseDocumentTransformer):
+class KeyBERTTransformer(BaseDocumentTransformer):
     """
     Add metadata to documents about keywords using `KeyBERT <https://maartengr.github.io/KeyBERT/>`_.
 
@@ -71,18 +72,14 @@ def __init__(
         *,
         batch_size: int = 8,
         metadata_key: str = "keywords",
-        model: str = "all-MiniLM-L6-v2",
+        model: str | KeyBERT = "all-MiniLM-L6-v2",
     ):
-        try:
-            import keybert  # type: ignore
-
-            self._kw_model = keybert.KeyBERT(model=model)
-        except ImportError:
-            raise ImportError(
-                "keybert is required for the KeybertLinkExtractor. "
-                "Please install it with `pip install keybert`."
-            ) from None
-
+        if isinstance(model, KeyBERT):
+            self._kw_model = model
+        elif isinstance(model, str):
+            self._kw_model = KeyBERT(model=model)
+        else:
+            raise ValueError(f"Invalid model: {model}")
         self._batch_size = batch_size
         self._metadata_key = metadata_key