Add prophet flavor (mlflow#4773)

* WIP build Signed-off-by: benwilson <benjamin.wilson@databricks.com> * Finish unit tests and address prophet build dependencies for serving Signed-off-by: benwilson <benjamin.wilson@databricks.com> * refactor example for prophet and add tracking validation test Signed-off-by: benwilson <benjamin.wilson@databricks.com> * adjusting build env Signed-off-by: benwilson <benjamin.wilson@databricks.com> * linting Signed-off-by: benwilson <benjamin.wilson@databricks.com> * Increase HTTP timeout to 90s. Disabled cloud storage HTTP timeout. (mlflow#4764) * Increase HTTP timeout to 120s. Disabled cloud storage HTTP timeout. * Enable auto dependency inference in spark flavor (mlflow#4759) * use infer_pip_requirements to spark.py Signed-off-by: harupy <hkawamura0130@gmail.com> * fix tests Signed-off-by: harupy <hkawamura0130@gmail.com> * fix tests Signed-off-by: harupy <hkawamura0130@gmail.com> * rename test Signed-off-by: harupy <hkawamura0130@gmail.com> * workaround for databricks Signed-off-by: harupy <hkawamura0130@gmail.com> * workaround for pyspark in databricks Signed-off-by: harupy <hkawamura0130@gmail.com> * rename test Signed-off-by: harupy <hkawamura0130@gmail.com> * address comments Signed-off-by: harupy <hkawamura0130@gmail.com> * better error message Signed-off-by: harupy <hkawamura0130@gmail.com> * fix error message Signed-off-by: harupy <hkawamura0130@gmail.com> * Fix `test_autolog_emits_warning_message_when_model_prediction_fails` (mlflow#4768) * Use AttributeError Signed-off-by: harupy <hkawamura0130@gmail.com> * comment Signed-off-by: harupy <hkawamura0130@gmail.com> * fix Signed-off-by: harupy <hkawamura0130@gmail.com> * use sanity_checking (mlflow#4767) Signed-off-by: harupy <hkawamura0130@gmail.com> * refactor: Extract the docker image building _build_image_from_context function (mlflow#4769) * refactor: Extract the docker image building _build_image_from_context function Signed-off-by: Alexey Volkov <alexey.volkov@ark-kun.com> * blacken Signed-off-by: harupy <hkawamura0130@gmail.com> Co-authored-by: harupy <hkawamura0130@gmail.com> * Fix autologging compatibility with Keras >= 2.6.0 (mlflow#4766) * Increase HTTP timeout to 90s. Disabled cloud storage HTTP timeout. (mlflow#4764) * Increase HTTP timeout to 120s. Disabled cloud storage HTTP timeout. Signed-off-by: dbczumar <corey.zumar@databricks.com> * Include keras conditionally Signed-off-by: dbczumar <corey.zumar@databricks.com> * Fixes Signed-off-by: dbczumar <corey.zumar@databricks.com> * Fix Signed-off-by: dbczumar <corey.zumar@databricks.com> * Fix root cause Signed-off-by: dbczumar <corey.zumar@databricks.com> * docstring Signed-off-by: dbczumar <corey.zumar@databricks.com> * Some test cases Signed-off-by: dbczumar <corey.zumar@databricks.com> * Some test cases Signed-off-by: dbczumar <corey.zumar@databricks.com> * Tests Signed-off-by: dbczumar <corey.zumar@databricks.com> * Format Signed-off-by: dbczumar <corey.zumar@databricks.com> * Test fixes Signed-off-by: dbczumar <corey.zumar@databricks.com> * Test fix 2 Signed-off-by: dbczumar <corey.zumar@databricks.com> * Remove keras change Signed-off-by: dbczumar <corey.zumar@databricks.com> * Use is Signed-off-by: dbczumar <corey.zumar@databricks.com> * Remove unused modules Signed-off-by: dbczumar <corey.zumar@databricks.com> * Use fixtures Signed-off-by: dbczumar <corey.zumar@databricks.com> * Docstring Signed-off-by: dbczumar <corey.zumar@databricks.com> * Fix fixtures Signed-off-by: dbczumar <corey.zumar@databricks.com> * Lint fixes Signed-off-by: dbczumar <corey.zumar@databricks.com> * Format Signed-off-by: dbczumar <corey.zumar@databricks.com> * Try preserve find module Signed-off-by: dbczumar <corey.zumar@databricks.com> * fix Signed-off-by: Weichen Xu <weichen.xu@databricks.com> * fix2 Signed-off-by: Weichen Xu <weichen.xu@databricks.com> * Simplify fluent test cases Signed-off-by: dbczumar <corey.zumar@databricks.com> * Format Signed-off-by: dbczumar <corey.zumar@databricks.com> * Tweaks, add a warning Signed-off-by: dbczumar <corey.zumar@databricks.com> * Test excludee Signed-off-by: dbczumar <corey.zumar@databricks.com> * Reverts Signed-off-by: dbczumar <corey.zumar@databricks.com> Co-authored-by: jinzhang21 <78067366+jinzhang21@users.noreply.github.com> Co-authored-by: Weichen Xu <weichen.xu@databricks.com> * Update spark support version to be 3.2 (mlflow#4770) * init Signed-off-by: Weichen Xu <weichen.xu@databricks.com> * Fix set matrix Signed-off-by: dbczumar <corey.zumar@databricks.com> * Fix Signed-off-by: dbczumar <corey.zumar@databricks.com> * Format Signed-off-by: dbczumar <corey.zumar@databricks.com> Co-authored-by: dbczumar <corey.zumar@databricks.com> * merge * merge * fix build errors * Changelog message tweak (mlflow#4777) * Tweak Signed-off-by: dbczumar <corey.zumar@databricks.com> * Lang Signed-off-by: dbczumar <corey.zumar@databricks.com> * Update CHANGELOG.rst * [ALL TESTS] Update (mlflow#4778) * Update MLflow version to 1.20.3 Signed-off-by: Jenkins <jenkins@databricks.com> * Update Signed-off-by: dbczumar <corey.zumar@databricks.com> Co-authored-by: Jenkins <jenkins@databricks.com> Co-authored-by: dbczumar <corey.zumar@databricks.com> * Fix small typo (mlflow#4772) Before: ``` 2021/09/03 18:53:45 WARNING mlflow.sklearn.utils: precision_score failed. The metric test_gold_precision_scorewill not be recorded. Metric error: '<' not supported between instances of 'float' and 'str' 2021/09/03 18:53:45 WARNING mlflow.sklearn.utils: recall_score failed. The metric test_gold_recall_scorewill not be recorded. Metric error: '<' not supported between instances of 'float' and 'str' 2021/09/03 18:53:45 WARNING mlflow.sklearn.utils: f1_score failed. The metric test_gold_f1_scorewill not be recorded. Metric error: '<' not supported between instances of 'float' and 'str' 2021/09/03 18:53:45 WARNING mlflow.sklearn.utils: accuracy_score failed. The metric test_gold_accuracy_scorewill not be recorded. Metric error: '<' not supported between instances of 'float' and 'str' ``` Signed-off-by: Louis Guitton <louisguitton93@gmail.com> * Push MLflow model to Sagemaker model registry (mlflow#4669) * Added push_sagemaker_model() api to enable push model from MLflow to Sagemaker model registry Signed-off-by: Jinni Gu <jinnigu@uw.edu> * Fixed the python doc for _find_transform_job() Signed-off-by: Jinni Gu <jinnigu@uw.edu> * Replace https protocol with s3:// protocol for the URL where Sagemaker model artifacts are stored in S3. Signed-off-by: Yiqing Wang <yiqing@wangemail.com> * rename push_sagemaker_model function add CLI function for push_model_to_sagemaker Signed-off-by: qtz123 <qiutingzhi1995@gmail.com> * reformat CLI for push_model_to_sagemaker Signed-off-by: qtz123 <qiutingzhi1995@gmail.com> * Used describe_model instead of list_models to check if the model exists. Signed-off-by: Yiqing Wang <yiqing@wangemail.com> * Renamed the _find_model method to _does_model_exist and limit the scope Signed-off-by: Yiqing Wang <yiqing@wangemail.com> * Fixed lint error and changed command to push model Signed-off-by: Yiqing Wang <yiqing@wangemail.com> Co-authored-by: Yiqing Wang <yiqing@wangemail.com> Co-authored-by: qtz123 <qiutingzhi1995@gmail.com> * Fix tensorflow dev cross test (mlflow#4780) * init Signed-off-by: Weichen Xu <weichen.xu@databricks.com> * dummy tf update Signed-off-by: Weichen Xu <weichen.xu@databricks.com> * Adding requirements for prophet examples Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * update conda versions for prophet Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * Fixing the prophet tests Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * PR fixes Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * Add notes on sub-dependencies Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * Updating docs for Prophet flavor Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * Resolve dependencies and add flavor tests Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> * update conda.yaml in examples for prophet Signed-off-by: Ben Wilson <benjamin.wilson@databricks.com> Co-authored-by: jinzhang21 <78067366+jinzhang21@users.noreply.github.com> Co-authored-by: Harutaka Kawamura <hkawamura0130@gmail.com> Co-authored-by: Alexey Volkov <alexey.volkov@ark-kun.com> Co-authored-by: dbczumar <39497902+dbczumar@users.noreply.github.com> Co-authored-by: Weichen Xu <weichen.xu@databricks.com> Co-authored-by: dbczumar <corey.zumar@databricks.com> Co-authored-by: mlflow-automation <61449322+mlflow-automation@users.noreply.github.com> Co-authored-by: Jenkins <jenkins@databricks.com> Co-authored-by: Louis Guitton <louisguitton@users.noreply.github.com> Co-authored-by: Jinni Gu <jinnigu@uw.edu> Co-authored-by: Yiqing Wang <yiqing@wangemail.com> Co-authored-by: qtz123 <qiutingzhi1995@gmail.com> Signed-off-by: Anjali Samad <samad.anjali.14@gmail.com>
anjsam1402 · Sep 14, 2021 · e5b486b · e5b486b
1 parent cb54575
commit e5b486b
Show file tree

Hide file tree

Showing 15 changed files with 779 additions and 276 deletions.
diff --git a/conftest.py b/conftest.py
@@ -94,6 +94,7 @@ def pytest_ignore_collect(path, config):
             "tests/models",
             "tests/shap",
             "tests/paddle",
+            "tests/prophet",
             "tests/utils/test_model_utils.py",
             # this test is included here because it imports many big libraries like tf, keras, etc
             "tests/tracking/fluent/test_fluent_autolog.py",

diff --git a/dev/extra-ml-requirements.txt b/dev/extra-ml-requirements.txt
@@ -42,3 +42,9 @@ pyspark
 shap
 # Required by mlflow.paddle
 paddlepaddle
+# Required by mlflow.prophet
+# NOTE: Prophet's whl build process will fail with dependencies not being present.
+#   Installation will default to setup.py in order to install correctly.
+#   To install in dev environment, ensure that gcc>=8 is installed to allow pystan
+#   to compile the model binaries. See: https://gcc.gnu.org/install/
+prophet
diff --git a/dev/run-python-flavor-tests.sh b/dev/run-python-flavor-tests.sh
@@ -12,10 +12,11 @@ pytest tests/pyfunc --large
 pytest tests/azureml --large
 pytest tests/utils/test_model_utils.py --large
 
-# TODO: Run tests for h2o, shap, and paddle in the cross-version-tests workflow
+# TODO: Run tests for h2o, shap, prophet, and paddle in the cross-version-tests workflow
 pytest tests/h2o --large
 pytest tests/shap --large
 pytest tests/paddle --large
+pytest tests/prophet --large
 
 pytest tests/tracking/fluent/test_fluent_autolog.py --large
 pytest tests/autologging --large

diff --git a/docs/source/models.rst b/docs/source/models.rst
@@ -763,6 +763,19 @@ on a ``statsmodels`` model.
 
 For more information, see :py:mod:`mlflow.statsmodels`.
 
+Prophet (``prophet``)
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+The ``prophet`` model flavor enables logging of `Prophet models
+<https://facebook.github.io/prophet/>`_ in MLflow format via the :py:func:`mlflow.prophet.save_model()`
+and :py:func:`mlflow.prophet.log_model()` methods.
+These methods also add the ``python_function`` flavor to the MLflow Models that they produce, allowing the
+models to be interpreted as generic Python functions for inference via
+:py:func:`mlflow.pyfunc.load_model()`. This loaded PyFunc model can only be scored with DataFrame input.
+You can also use the :py:func:`mlflow.prophet.load_model()`
+method to load MLflow Models with the ``prophet`` model flavor in native prophet format.
+
+For more information, see :py:mod:`mlflow.prophet`.
+
 Model Customization
 -------------------
 

diff --git a/docs/source/python_api/mlflow.prophet.rst b/docs/source/python_api/mlflow.prophet.rst
@@ -0,0 +1,7 @@
+mlflow.prophet
+==================
+
+.. automodule:: mlflow.prophet
+    :members:
+    :undoc-members:
+    :show-inheritance:
diff --git a/docs/source/tutorials-and-examples/index.rst b/docs/source/tutorials-and-examples/index.rst
@@ -41,6 +41,8 @@ Below, you can find a number of tutorials and examples for various MLflow use ca
 
   - `SHAP <https://github.com/mlflow/mlflow/tree/master/examples/shap>`_
 
+  - `Prophet <https://github.com/mlflow/mlflow/tree/master/examples/prophet>`_
+
   - scikit-learn
 
     + `Diabetes example <https://github.com/mlflow/mlflow/tree/master/examples/sklearn_elasticnet_diabetes>`_

diff --git a/examples/prophet/MLproject b/examples/prophet/MLproject
@@ -1,14 +1,7 @@
-name: tutorial
+name: prophet_example
 
 conda_env: conda.yaml
 
 entry_points:
   main:
-    parameters:
-      csv_url: {type: uri, default: 'https://raw.githubusercontent.com/facebook/prophet/e21a05f4f9290649255a2a306855e8b4620816d7/examples/example_wp_log_peyton_manning.csv'}
-      rolling_window: {type: float, default: 0.1}
-    command: "python train.py {csv_url} {rolling_window}"
-  predict:
-    parameters:
-      model_uri: {type: uri}
-    command: "python predict.py {model_uri}"
+    command: "python train.py"
diff --git a/examples/prophet/conda.yaml b/examples/prophet/conda.yaml
@@ -1,9 +1,9 @@
-name: tutorial
+name: prophet-example
 channels:
   - conda-forge
 dependencies:
   - python=3.6
   - pip
   - pip:
-    - mlflow>=1.0
-    - fbprophet<=0.7
+    - prophet>=1.0.1
+
diff --git a/examples/prophet/predict.py b/examples/prophet/predict.py
diff --git a/examples/prophet/train.ipynb b/examples/prophet/train.ipynb