Document check copies (#25291)

* Document check copies better and add tests

* Include header in check for copies

* Manual fixes

* Try autofix

* Fixes

* Clean tests

* Finalize doc

* Remove debug print

* More fixes

docs/source/en/add_new_model.md

@@ -101,7 +101,7 @@ own regarding how code should be written :-)
 1. The forward pass of your model should be fully written in the modeling file while being fully independent of other
    models in the library. If you want to reuse a block from another model, copy the code and paste it with a
    `# Copied from` comment on top (see [here](https://github.com/huggingface/transformers/blob/v4.17.0/src/transformers/models/roberta/modeling_roberta.py#L160)
-   for a good example).
+   for a good example and [there](pr_checks#check-copies) for more documentation on Copied from). 
 2. The code should be fully understandable, even by a non-native English speaker. This means you should pick
    descriptive variable names and avoid abbreviations. As an example, `activation` is preferred to `act`.
    One-letter variable names are strongly discouraged unless it's an index in a for loop.

docs/source/en/pr_checks.md

@@ -142,3 +142,58 @@ Additional checks concern PRs that add new models, mainly that:
 - All checkpoints used actually exist on the Hub
 
 -->
+
+### Check copies
+
+Since the Transformers library is very opinionated with respect to model code, and each model should fully be implemented in a single file without relying on other models, we have added a mechanism that checks whether a copy of the code of a layer of a given model stays consistent with the original. This way, when there is a bug fix, we can see all other impacted models and choose to trickle down the modification or break the copy.
+
+<Tip>
+
+If a file is a full copy of another file, you should register it in the constant `FULL_COPIES` of `utils/check_copies.py`.
+
+</Tip>
+
+This mechanism relies on comments of the form `# Copied from xxx`. The `xxx` should contain the whole path to the class of function which is being copied below. For instance, `RobertaSelfOutput` is a direct copy of the `BertSelfOutput` class, so you can see [here](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L289) it has a comment:
+
+```py
+# Copied from transformers.models.bert.modeling_bert.BertSelfOutput
+```
+
+Note that instead of applying this to a whole class, you can apply it to the relevant methods that are copied from. For instance [here](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L598) you can see how `RobertaPreTrainedModel._init_weights` is copied from the same method in `BertPreTrainedModel` with the comment:
+
+```py
+# Copied from transformers.models.bert.modeling_bert.BertPreTrainedModel._init_weights
+```
+
+Sometimes the copy is exactly the same except for names: for instance in `RobertaAttention`, we use `RobertaSelfAttention` insted of `BertSelfAttention` but other than that, the code is exactly the same. This is why `# Copied from` supports simple string replacements with the follwoing syntax: `Copied from xxx with foo->bar`. This means the code is copied with all instances of `foo` being replaced by `bar`. You can see how it used [here](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L304C1-L304C86) in `RobertaAttention` with the comment:
+
+```py
+# Copied from transformers.models.bert.modeling_bert.BertAttention with Bert->Roberta
+```
+
+Note that there shouldn't be any spaces around the arrow (unless that space is part of the pattern to replace of course).
+
+You can add several patterns separated by a comma. For instance here `CamemberForMaskedLM` is a direct copy of `RobertaForMaskedLM` with two replacements: `Roberta` to `Camembert` and `ROBERTA` to `CAMEMBERT`. You can see [here](https://github.com/huggingface/transformers/blob/15082a9dc6950ecae63a0d3e5060b2fc7f15050a/src/transformers/models/camembert/modeling_camembert.py#L929) this is done with the comment:
+
+```py
+# Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM with Roberta->Camembert, ROBERTA->CAMEMBERT
+```
+
+If the order matters (because one of the replacements might conflict with a previous one), the replacements are executed from left to right.
+
+<Tip>
+
+If the replacements change the formatting (if you replace a short name by a very long name for instance), the copy is checked after applying the auto-formatter.
+
+</Tip>
+
+Another way when the patterns are just different casings of the same replacement (with an uppercased and a lowercased variants) is just to add the option `all-casing`. [Here](https://github.com/huggingface/transformers/blob/15082a9dc6950ecae63a0d3e5060b2fc7f15050a/src/transformers/models/mobilebert/modeling_mobilebert.py#L1237) is an example in `MobileBertForSequenceClassification` with the comment:
+
+```py
+# Copied from transformers.models.bert.modeling_bert.BertForSequenceClassification with Bert->MobileBert all-casing
+```
+
+In this case, the code is copied from `BertForSequenceClassification` by replacing:
+- `Bert` by `MobileBert` (for instance when using `MobileBertModel` in the init)
+- `bert` by `mobilebert` (for instance when defining `self.mobilebert`)
+- `BERT` by `MOBILEBERT` (in the constant `MOBILEBERT_INPUTS_DOCSTRING`)