chore: Change trafilatura dependency to use lazy import (#7809)

* Change trafilatura dependency to use lazy import

* Add release notes

haystack/components/converters/html.py

@@ -6,14 +6,16 @@ import warnings
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Union
 
-from trafilatura import extract
-
 from haystack import Document, component, default_from_dict, default_to_dict, logging
 from haystack.components.converters.utils import get_bytestream_from_source, normalize_metadata
 from haystack.dataclasses import ByteStream
+from haystack.lazy_imports import LazyImport
 
 logger = logging.getLogger(__name__)
 
+with LazyImport("Run 'pip install trafilatura'") as trafilatura_import:
+    from trafilatura import extract
+
 
 @component
 class HTMLToDocument:
@@ -49,6 +51,7 @@ class HTMLToDocument:
             are passed to the underlying Trafilatura `extract` function. For the full list of available arguments, see
             the [Trafilatura documentation](https://trafilatura.readthedocs.io/en/latest/corefunctions.html#extract).
         """
+        trafilatura_import.check()
         if extractor_type is not None:
             warnings.warn(
                 "The `extractor_type` parameter is ignored and will be removed in Haystack 2.4.0. "

pyproject.toml

@@ -57,7 +57,6 @@ dependencies = [
   "more-itertools",  # TextDocumentSplitter
   "networkx", # Pipeline graphs
   "typing_extensions>=4.7", # typing support for Python 3.8
-  "trafilatura", # Fulltext extraction from HTML pages
   "requests",
   "numpy",
   "python-dateutil",
@@ -117,6 +116,7 @@ extra-dependencies = [
   "langdetect",  # TextLanguageRouter and DocumentLanguageClassifier
   "sentence-transformers>=2.2.0",  # SentenceTransformersTextEmbedder and SentenceTransformersDocumentEmbedder
   "openai-whisper>=20231106",  # LocalWhisperTranscriber
+  "trafilatura", # Fulltext extraction from HTML pages
 
   # OpenAPI
   "jsonref",  # OpenAPIServiceConnector, OpenAPIServiceToFunctions

releasenotes/notes/change-trafilatura-dependency-575985a2271c8370.yaml

@@ -0,0 +1,7 @@
+---
+upgrade:
+  - |
+    `trafilatura` must now be manually installed with `pip install trafilatura` to use the `HTMLToDocument` Component.
+enhancements:
+  - |
+    Remove `trafilatura` as direct dependency and make it a lazily imported one