fix typos (#837)

Most of my remarks were in the custom component page. But I see that
this page has been removed in the meantime.

The image I removed had a black background which was not very visible on
the documentation. But since custom component page is removed it is not
used anymore. Hence the removal.

Your welcome for my sick contribution.

docs/components/component_spec.md

@@ -85,7 +85,7 @@ dataset.
 
 Each field defines the expected data type, which should match the
 [types defined by Fondant](https://github.com/ml6team/fondant/blob/main/src/fondant/core/schema.py),
-that correespond to [Arrow data types](https://arrow.apache.org/docs/python/api/datatypes.html).
+that correspond to [Arrow data types](https://arrow.apache.org/docs/python/api/datatypes.html).
 
 Note that you can always map a field from your dataset with a different name to a specific field name expected by the
 component provided they have the same data type. For example, suppose we have a component spec that

docs/components/containerized_components.md

@@ -12,7 +12,7 @@ with a simpler [lightweight components](../components/lightweight_components.md)
 
 To make sure containerized components are reusable, they should implement a single logical data
 processing
-step (like captioning images or removing Personal Identifiable Information [PII] from text.)
+step (like captioning images or removing Personal Identifiable Information [PII] from text).
 If a component grows too large, consider splitting it into multiple separate components each
 tackling one logical part.
 
@@ -38,7 +38,7 @@ script.
 
 Note that the `main.py` script can be split up into several Python scripts in case it would become
 prohibitively long. See the
-[prompt based LAION retrieval component](https://github.com/ml6team/fondant/tree/main/components/prompt_based_laion_retrieval/src)
+[prompt based LAION retrieval component](https://github.com/ml6team/fondant/tree/main/components/retrieve_laion_by_prompt/src)
 as an example: the CLIP client itself is defined in a separate script called `clip_client`,
 which is then imported in the `main.py` script.