Add details about pretrained pipeline design

website/docs/models/index.md

@@ -4,6 +4,7 @@ teaser: Downloadable trained pipelines and weights for spaCy
 menu:
   - ['Quickstart', 'quickstart']
   - ['Conventions', 'conventions']
+  - ['Pipeline Design', 'design']
 ---
 
 <!-- TODO: include interactive demo -->
@@ -53,3 +54,146 @@ For a detailed compatibility overview, see the
 [`compatibility.json`](https://github.com/explosion/spacy-models/tree/master/compatibility.json).
 This is also the source of spaCy's internal compatibility check, performed when
 you run the [`download`](/api/cli#download) command.
+
+## Pretrained pipeline design {#design}
+
+The spaCy v3 pretrained pipelines are designed to be efficient and configurable.
+For example, multiple components can share a common "token-to-vector" model and
+it's easy to swap out or disable the lemmatizer. The pipelines are designed to
+be efficient in terms of speed and size and work well when the pipeline is run
+in full.
+
+When modifying a pretrained v3 pipeline, it's important to understand how the
+components **depend on** each other. Unlike spaCy v2, where the `tagger`,
+`parser` and `ner` components were all independent, some v3 components depend on
+earlier components in the pipeline. As a result, disabling or reordering
+components can affect the annotation quality or lead to warnings and errors.
+
+Main changes from spaCy v2 models:
+
+- The [`Tok2Vec`](/api/tok2vec) component may be a separate, shared component. A
+  component like a tagger or parser can
+  [listen](/api/architectures#Tok2VecListener) to an earlier `tok2vec` or
+  `transformer` rather than having its own separate tok2vec layer.
+- Rule-based exceptions move from individual components to the
+  `attribute_ruler`. Lemma and POS exceptions move from the tokenizer exceptions
+  to the attribute ruler and the tag map and morph rules move from the tagger to
+  the attribute ruler.
+- The lemmatizer tables and processing move from the vocab and tagger to a
+  separate `lemmatizer` component.
+
+### CNN/CPU pipeline design
+
+In the `sm`/`md`/`lg` models:
+
+- The `tagger`, `morphologizer` and `parser` components listen to the `tok2vec`
+  component.
+- The `attribute_ruler` maps `token.tag` to `token.pos` if there is no
+  `morphologizer`. The `attribute_ruler` additionally makes sure whitespace is
+  tagged consistently and copies `token.pos` to `token.tag` if there is no
+  tagger. For English, the attribute ruler can improve its mapping from
+  `token.tag` to `token.pos` if dependency parses from a `parser` are present,
+  but the parser is not required.
+- The rule-based `lemmatizer` (Dutch, English, French, Greek, Macedonian,
+  Norwegian and Spanish) requires `token.pos` annotation from either
+  `tagger`+`attribute_ruler` or `morphologizer`.
+- The `ner` component is independent with its own internal tok2vec layer.
+
+<!-- TODO: pretty diagram -->
+
+### Transformer pipeline design
+
+In the tranformer (`trf`) models, the `tagger`, `parser` and `ner` (if present)
+all listen to the `transformer` component. The `attribute_ruler` and
+`lemmatizer` have the same configuration as in the CNN models.
+
+<!-- TODO: pretty diagram -->
+
+### Modifying the default pipeline
+
+For faster processing, you may only want to run a subset of the components in a
+pretrained pipeline. The `disable` and `exclude` arguments to
+[`spacy.load`](/api/top-level#spacy.load) let you control which components are
+loaded and run. Disabled components are loaded in the background so it's
+possible to reenable them in the same pipeline in the future with
+[`nlp.enable_pipe`](/api/language/#enable_pipe). To skip loading a component
+completely, use `exclude` instead of `disable`.
+
+#### Disable part-of-speech tagging and lemmatization
+
+To disable part-of-speech tagging and lemmatization, disable the `tagger`,
+`morphologizer`, `attribute_ruler` and `lemmatizer` components.
+
+```python
+# Note: English doesn't include a morphologizer
+nlp = spacy.load("en_core_web_sm", disable=["tagger", "attribute_ruler", "lemmatizer"])
+nlp = spacy.load("en_core_web_trf", disable=["tagger", "attribute_ruler", "lemmatizer"])
+```
+
+<Infobox variant="warning" title="Rule-based lemmatizers require Token.pos">
+
+The lemmatizer depends on `tagger`+`attribute_ruler` or `morphologizer` for
+Dutch, English, French, Greek, Macedonian, Norwegian and Spanish. If you disable
+any of these components, you'll see lemmatizer warnings unless the lemmatizer is
+also disabled.
+
+</Infobox>
+
+#### Use senter rather than parser for fast sentence segmentation
+
+If you need fast sentence segmentation without dependency parses, disable the
+`parser` use the `senter` component instead:
+
+```python
+nlp = spacy.load("en_core_web_sm")
+nlp.disable_pipe("parser")
+nlp.enable_pipe("senter")
+```
+
+The `senter` component is ~10&times; faster than the parser and more accurate
+than the rule-based `sentencizer`.
+
+#### Switch from rule-based to lookup lemmatization
+
+For the Dutch, English, French, Greek, Macedonian, Norwegian and Spanish
+pipelines, you can switch from the default rule-based lemmatizer to a lookup
+lemmatizer:
+
+```python
+# Requirements: pip install spacy-lookups-data
+nlp = spacy.load("en_core_web_sm")
+nlp.remove_pipe("lemmatizer")
+nlp.add_pipe("lemmatizer", config={"mode": "lookup"}).initialize()
+```
+
+#### Disable everything except NER
+
+For the non-transformer models, the `ner` component is independent, so you can
+disable everything else:
+
+```python
+nlp = spacy.load("en_core_web_sm", disable=["tok2vec", "tagger", "parser", "attribute_ruler", "lemmatizer"])
+```
+
+In the transformer models, `ner` listens to the `transformer` layer, so you can
+disable all components related tagging, parsing, and lemmatization.
+
+```python
+nlp = spacy.load("en_core_web_trf", disable=["tagger", "parser", "attribute_ruler", "lemmatizer"])
+```
+
+#### Move NER to the end of the pipeline
+
+For access to `POS` and `LEMMA` features in an `entity_ruler`, move `ner` to the
+end of the pipeline after `attribute_ruler` and `lemmatizer`:
+
+```python
+# load without NER
+nlp = spacy.load("en_core_web_sm", exclude=["ner"])
+
+# source NER from the same pipeline package as the last component
+nlp.add_pipe("ner", source=spacy.load("en_core_web_sm"))
+
+# insert the entity ruler
+nlp.add_pipe("entity_ruler", before="ner")
+```