merge dev for release 0.9.7

.github/CONTRIBUTING.md

@@ -7,16 +7,16 @@ We love your input! We want to make contributing to this project as easy and tra
 - Proposing new features
 - Becoming a maintainer
 
-## We Develop with Github
-We use github to host code, to track issues and feature requests, as well as accept pull requests.
+## We Develop with GitHub
+We use GitHub to host code, to track issues and feature requests, as well as accept pull requests.
 
-## We Use [Github Flow](https://guides.github.com/introduction/flow/index.html), So All Code Changes Happen Through Pull Requests
+## We Use [GitHub Flow](https://guides.github.com/introduction/flow/index.html), So All Code Changes Happen Through Pull Requests
 Pull requests are the best way to propose changes to the codebase. We actively welcome your pull requests:
 
 1. Fork the repo and create your branch from the **`dev` branch**.
 2. If you've added code that should be tested, you **must** ensure it is properly tested.
 3. If you've changed APIs, update the documentation.
-4. Ensure the Travis test suite passes.
+4. Ensure the CI/CD test suite passes.
 5. Make sure your code lints.
 6. Submit that pull request!
 

.github/workflows/docs.yml

@@ -21,6 +21,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
         run: |
+          python -m pip install --upgrade pip
           pip install -Ur requirements.txt
           pip install -Ur docs/requirements.txt
           pip install -e .

.github/workflows/linting.yml

@@ -17,6 +17,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
         run: |
+          python -m pip install --upgrade pip
           pip install -Ur requirements.txt
           pip install -Ur styling_requirements.txt
           pip install -Ur tests/requirements.txt

.github/workflows/tests.yml

@@ -17,6 +17,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
         run: |
+          python -m pip install --upgrade pip
           pip install -Ur requirements.txt
           pip install -Ur tests/requirements.txt
           python setup.py develop
@@ -38,6 +39,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
         run: |
+          python -m pip install --upgrade pip
           pip install -Ur requirements.txt
           pip install -Ur tests/requirements.txt
           python setup.py develop

.gitignore

@@ -144,3 +144,5 @@ deepparse/version.py
 *.ckpt
 
 *mlruns/
+
+*model/

.release/bpemb.version

@@ -0,0 +1 @@
+aa32fa918494b461202157c57734c374

.release/bpemb_attention.version

@@ -0,0 +1 @@
+cfb190902476376573591c0ec6f91ece

.release/fasttext.version

@@ -0,0 +1 @@
+f67a0517c70a314bdde0b8440f21139d

.release/fasttext_attention.version

@@ -0,0 +1 @@
+a2b688bdfa2aa7c009bb7d980e352978

.release/model_version_release.md

@@ -0,0 +1,5 @@
+# How to Create a New Model's Version
+
+1. `md5sum <model.ckpt> > model.version`
+2. Remove the model.cpkt text in `model.version` file
+3. Update latests BPEMB and FastText hash in `tests/test_tools.py`

CHANGELOG.md

@@ -313,4 +313,17 @@
   suggested in the [documentation](https://pytorch.org/tutorials//intermediate/torch_compile_tutorial.html). It
   increases the performance by about 1/100.
 
+## 0.9.7
+
+- New models release with more meta-data.
+- Add a feature to use an AddressParser from a URI.
+- Add a feature to upload the trained model to a URI.
+- Add an example of how to use URI for parsing from and uploading to.
+- Improve error handling of `path_to_retrain_model`.
+- Bug-fix pre-processor error.
+- Add verbose override and improve verbosity handling in retrain.
+- Bug-fix the broken FastText installation using `fasttext-wheel` instead of `fasttext` (
+  see [here](https://github.com/facebookresearch/fastText/issues/512#issuecomment-1534519551)
+  and [here](https://github.com/facebookresearch/fastText/pull/1292)).
+
 ## dev

deepparse/__init__.py

@@ -2,4 +2,4 @@
 from .fasttext_tools import *
 from .tools import *
 from .version import __version__
-from .weights_init import *
+from .weights_tools import *

deepparse/cli/parse.py

@@ -50,7 +50,7 @@ def main(args=None) -> None:
 
     .. code-block:: sh
 
-        parse fasttext ./dataset.csv parsed_address.pckl --path_to_retrained_model ./path
+        parse fasttext ./dataset.csv parsed_address.pckl --path_to_model_weights ./path
 
     """
     if args is None:  # pragma: no cover

deepparse/cli/parser_arguments_adder.py

@@ -108,7 +108,7 @@ def add_batch_size_arg(parser: ArgumentParser) -> None:
 def add_path_to_retrained_model_arg(parser: ArgumentParser) -> None:
     parser.add_argument(
         "--path_to_retrained_model",
-        help=wrap("A path to a retrained model to use for testing."),
+        help=wrap("A path to a retrained model to use. It can be an S3-URI."),
         type=str,
         default=None,
     )

deepparse/network/decoder.py

@@ -6,7 +6,7 @@
 import torch
 from torch import nn
 
-from ..weights_init import weights_init
+from .. import weights_init
 
 
 class Decoder(nn.Module):

deepparse/network/encoder.py

@@ -7,7 +7,7 @@
 from torch import nn
 from torch.nn.utils.rnn import pack_padded_sequence, pad_packed_sequence
 
-from ..weights_init import weights_init
+from .. import weights_init
 
 
 class Encoder(nn.Module):

deepparse/network/seq2seq.py

@@ -4,14 +4,14 @@
 import random
 import warnings
 from abc import ABC
-from collections import OrderedDict
 from typing import Tuple, Union, List
 
 import torch
 from torch import nn
 
 from .decoder import Decoder
 from .encoder import Encoder
+from .. import handle_weights_upload
 from ..tools import download_weights, latest_version
 
 
@@ -113,20 +113,21 @@ def _load_pre_trained_weights(self, model_type: str, cache_dir: str, offline: bo
                     )
                 download_weights(model_type, cache_dir, verbose=self.verbose)
 
-        all_layers_params = torch.load(model_path, map_location=self.device)
-        self.load_state_dict(all_layers_params)
+        self._load_weights(path_to_model_torch_archive=model_path)
 
-    def _load_weights(self, path_to_retrained_model: str) -> None:
+    def _load_weights(self, path_to_model_torch_archive: str) -> None:
         """
         Method to load (into the network) the weights.
 
         Args:
-            path_to_retrained_model (str): The path to the fine-tuned model.
+            path_to_model_torch_archive (str): The path to the fine-tuned model Torch archive.
         """
-        all_layers_params = torch.load(path_to_retrained_model, map_location=self.device)
-        if isinstance(all_layers_params, dict) and not isinstance(all_layers_params, OrderedDict):
-            # Case where we have a retrained model with a different tagging space
-            all_layers_params = all_layers_params.get("address_tagger_model")
+        all_layers_params = handle_weights_upload(
+            path_to_model_to_upload=path_to_model_torch_archive, device=self.device
+        )
+
+        # All the time, our torch archive include meta-data along with the model weights
+        all_layers_params = all_layers_params.get("address_tagger_model")
         self.load_state_dict(all_layers_params)
 
     def _encoder_step(self, to_predict: torch.Tensor, lengths: List, batch_size: int) -> Tuple: