naming and consistency (#174)

* naming and consistency

* update sample tests and add backwards compatability tests

* predict tests

* partial docs update

* pankaj review updates

Author: philipkiely-baseten

README.md

@@ -68,15 +68,15 @@ rfc = RandomForestClassifier()
 rfc.fit(data_x, data_y)
 
 # Create the Truss (serializing & packaging model)
-tr = truss.mk_truss(rfc, target_directory="iris_rfc_truss")
+tr = truss.create(rfc, target_directory="iris_rfc_truss")
 
 # Serve a prediction from the model
-tr.server_predict({"inputs": [[0, 0, 0, 0]]})
+tr.predict({"inputs": [[0, 0, 0, 0]]})
 ```
 
 ### Package your model
 
-The `truss.mk_truss()` command can be used with any supported framework:
+The `truss.create()` command can be used with any supported framework:
 
 * [Hugging Face](https://truss.baseten.co/create/huggingface)
 * [LightGBM](https://truss.baseten.co/create/lightgbm)

docs/create/huggingface.md

@@ -29,12 +29,12 @@ model = pipeline('fill-mask', model='bert-base-uncased')
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="huggingface_truss")
+tr = create(model, target_directory="huggingface_truss")
 ```
 
 Check the target directory to see your new Truss!

docs/create/lightgbm.md

@@ -51,12 +51,12 @@ model = lgb.train(params=params, train_set=train, valid_sets=test)
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="lightgbm_truss")
+tr = create(model, target_directory="lightgbm_truss")
 ```
 
 Check the target directory to see your new Truss!

docs/create/manual.md

@@ -16,7 +16,7 @@ To familiarize yourself with the structure of Truss, review the [structure refer
 
 ### Adding the model binary
 
-First, you'll need to add a model binary to your new Truss. On supported frameworks, this is provided automatically by the `mk_truss` command. For a custom Truss, it can come from many sources, such as:
+First, you'll need to add a model binary to your new Truss. On supported frameworks, this is provided automatically by the `create` command. For a custom Truss, it can come from many sources, such as:
 
 * Pickling your model
 * Serializing your model
@@ -46,7 +46,7 @@ Also, your model gets access to certain values, including the `config.yaml` file
 
 ## Example code
 
-While XGBoost is a supported framework — you can make a Truss from an XGBoost model with `mk_truss` — we'll use the manual method here for demonstration.
+While XGBoost is a supported framework — you can make a Truss from an XGBoost model with `create` — we'll use the manual method here for demonstration.
 
 If you haven't already, create a Truss by running:
 

docs/create/mlflow.md

@@ -35,14 +35,14 @@ with mlflow.start_run():
 
 ### Create a Truss
 
-Truss uses MLflow's [pyfunc](https://www.mlflow.org/docs/latest/python_api/mlflow.pyfunc.html) module in the packaging process. Once you have loaded the model, use the `mk_truss` command to package your model into a Truss.
+Truss uses MLflow's [pyfunc](https://www.mlflow.org/docs/latest/python_api/mlflow.pyfunc.html) module in the packaging process. Once you have loaded the model, use the `create` command to package your model into a Truss.
 
 ```python
 import os
 import truss
 
 model = mlflow.pyfunc.load_model(MODEL_URI)
-tr = truss.mk_truss(model, target_directory="./mlflow_truss")
+tr = truss.create(model, target_directory="./mlflow_truss")
 ```
 
 Check the target directory to see your new Truss!
@@ -53,7 +53,7 @@ To get a prediction from the Truss, try running:
 
 ```python
 data = np.array([-4, 1, 0, 10, -2, 1]).reshape(-1, 1)
-predictions = tr.server_predict({"inputs": data})
+predictions = tr.predict({"inputs": data})
 print(predictions)
 ```
 

docs/create/pytorch.md

@@ -143,12 +143,12 @@ model, _ , _ = train_the_model()
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="pytorch_truss")
+tr = create(model, target_directory="pytorch_truss")
 ```
 
 Check the target directory to see your new Truss!
@@ -165,7 +165,7 @@ inputs = datasets.MNIST("../data", train=False, transform=transform)
 dataset = torch.utils.data.DataLoader(inputs, batch_size=1)
 
 import numpy as np
-print(tr.server_predict({"inputs": np.array(next(iter(dataset))[0])}))
+print(tr.predict({"inputs": np.array(next(iter(dataset))[0])}))
 ```
 
 For information on running the Truss locally, see [local development](../develop/localhost.md).

docs/create/sklearn.md

@@ -30,12 +30,12 @@ model.fit(data_x, data_y)
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="sklearn_truss")
+tr = create(model, target_directory="sklearn_truss")
 ```
 
 Check the target directory to see your new Truss!

docs/create/tensorflow.md

@@ -32,12 +32,12 @@ model = tf.keras.applications.ResNet50V2(
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="tensorflow_truss")
+tr = create(model, target_directory="tensorflow_truss")
 ```
 
 Check the target directory to see your new Truss!
@@ -80,7 +80,7 @@ def postprocess(predictions, k=5):
 With these functions in place, you can invoke the model and pass it a URL, as in:
 
 ```python
-tr.server_predict({"inputs": "https://github.com/pytorch/hub/raw/master/images/dog.jpg"})
+tr.predict({"inputs": "https://github.com/pytorch/hub/raw/master/images/dog.jpg"})
 ```
 
 For information on running the Truss locally, see [local development](../develop/localhost.md).

docs/create/xgboost.md

@@ -46,12 +46,12 @@ model = xgb.train(params,
 
 ### Create a Truss
 
-Use the `mk_truss` command to package your model into a Truss.
+Use the `create` command to package your model into a Truss.
 
 ```python
-from truss import mk_truss
+from truss import create
 
-tr = mk_truss(model, target_directory="xgboost_truss")
+tr = create(model, target_directory="xgboost_truss")
 ```
 
 Check the target directory to see your new Truss!

docs/deploy/baseten.md

@@ -15,7 +15,7 @@ pip install --upgrade baseten
 
 {% hint style="info" %}
 
-If your model is already in memory (you created it with `mk_truss`), you can skip loading it into memory from the directory.
+If your model is already in memory (you created it with `create`), you can skip loading it into memory from the directory.
 
 {% endhint %}
 
@@ -24,7 +24,7 @@ Before deploying your Truss, you may need to load it into memory in a Jupyter no
 ```python
 import truss
 
-my_truss = truss.from_directory("my_truss_lives_here")
+my_truss = truss.load("my_truss_lives_here")
 ```
 
 Once your Truss is in memory, simply run the following:

docs/develop/localhost.md

@@ -21,7 +21,7 @@ There are two ways to interface with a Truss locally. The first is via [the Pyth
 When interacting with your Truss via the Python client, the first thing is to make sure it is in-memory. If you just created the Truss, it'll be in memory already, but if not, you'll need to load it with the following command:
 
 ```python
-tr = truss.from_directory("path_to_my_truss")
+tr = truss.load("path_to_my_truss")
 ```
 
 From there, you can invoke the Truss to serve the model in your Python environment. Just run:
@@ -112,7 +112,7 @@ Unlike Docker image, this mechanism requires that you already have the right Pyt
 In the Python environment, get a prediction without Docker by running:
 
 ```python
-tr.server_predict({"inputs": [[0, 0, 0, 0]]})
+tr.predict({"inputs": [[0, 0, 0, 0]]})
 ```
 
 Or in the command line, run:

docs/e2e.md

@@ -22,14 +22,14 @@ If you want to follow this tutorial for a particular framework and don't have a
 
 ## Step 1: Create a Truss
 
-Truss works across model frameworks, and the most common model frameworks are supported with the one-line `mk_truss` command. Click the framework you built your model in to see specific packaging instructions for that format.
+Truss works across model frameworks, and the most common model frameworks are supported with the one-line `create` command. Click the framework you built your model in to see specific packaging instructions for that format.
 
 For the following formats, if you have an in-memory trained model object `model`, just call the following:
 
 ```python
-from truss import mk_truss
+from truss import create
 
-mk_truss(model, target_directory="my_truss")
+create(model, target_directory="my_truss")
 ```
 
 Supported frameworks:

docs/notebooks/aws_example.ipynb

@@ -49,7 +49,7 @@
    "outputs": [],
    "source": [
     "# Make scaffold \n",
-    "scaffold = truss.mk_truss(rfc_model, target_directory='test_rfc_1')\n",
+    "scaffold = truss.create(rfc_model, target_directory='test_rfc_1')\n",
     "# This will produce a folder `test_rfc_1/` relative to your current directory. \n",
     "# Now we'll use Truss to build a Docker image on our local system\n",
     "scaffold.build_docker_image()"

docs/notebooks/huggingface_example.ipynb

@@ -44,9 +44,9 @@
       },
       "outputs": [],
       "source": [
-        "from truss import mk_truss\n",
+        "from truss import create\n",
         "\n",
-        "tr = mk_truss(model, target_directory=\"huggingface_truss\")"
+        "tr = create(model, target_directory=\"huggingface_truss\")"
       ]
     },
     {
@@ -58,7 +58,7 @@
       "outputs": [],
       "source": [
         "# Serve a prediction from the model\n",
-        "tr.server_predict({\"inputs\": [\"Donatello is a teenage mutant [MASK] turtle\"]})"
+        "tr.predict({\"inputs\": [\"Donatello is a teenage mutant [MASK] turtle\"]})"
       ]
     }
   ],

docs/notebooks/lightgbm_example.ipynb

@@ -68,9 +68,9 @@
       },
       "outputs": [],
       "source": [
-        "from truss import mk_truss\n",
+        "from truss import create\n",
         "\n",
-        "tr = mk_truss(model, target_directory=\"lightgbm_truss\")"
+        "tr = create(model, target_directory=\"lightgbm_truss\")"
       ]
     },
     {
@@ -81,7 +81,7 @@
       },
       "outputs": [],
       "source": [
-        "tr.server_predict({\"inputs\": [[0, 0, 0, 0, 0, 0]]})"
+        "tr.predict({\"inputs\": [[0, 0, 0, 0, 0, 0]]})"
       ]
     }
   ],

docs/notebooks/mlflow_example.ipynb

@@ -85,7 +85,7 @@
         "import truss\n",
         "\n",
         "model = mlflow.pyfunc.load_model(MODEL_URI)\n",
-        "tr = truss.mk_truss(model, target_directory=\"./mlflow_truss\")"
+        "tr = truss.create(model, target_directory=\"./mlflow_truss\")"
       ]
     },
     {
@@ -99,7 +99,7 @@
         "# Invoke the MLflow model\n",
         "\n",
         "data = np.array([-4, 1, 0, 10, -2, 1]).reshape(-1, 1)\n",
-        "predictions = tr.server_predict({\"inputs\": data})\n",
+        "predictions = tr.predict({\"inputs\": data})\n",
         "print(predictions)"
       ]
     }

docs/notebooks/sklearn_example.ipynb

@@ -60,7 +60,7 @@
       "outputs": [],
       "source": [
         "# Create the Truss (serializing & packaging model)\n",
-        "tr = truss.mk_truss(rfc, \"iris-rfc\")"
+        "tr = truss.create(rfc, \"iris-rfc\")"
       ]
     },
     {
@@ -72,7 +72,7 @@
       "outputs": [],
       "source": [
         "# Serve a prediction from the model\n",
-        "tr.server_predict({\"inputs\": [[0, 0, 0, 0]]})"
+        "tr.predict({\"inputs\": [[0, 0, 0, 0]]})"
       ]
     }
   ],

docs/notebooks/tensorflow_example.ipynb

@@ -49,10 +49,10 @@
       },
       "outputs": [],
       "source": [
-        "from truss import mk_truss\n",
+        "from truss import create\n",
         "\n",
         "# Create the Truss (serializing & packaging model)\n",
-        "tr = mk_truss(model, target_directory=\"tensorflow_truss\")"
+        "tr = create(model, target_directory=\"tensorflow_truss\")"
       ]
     },
     {
@@ -102,7 +102,7 @@
       "outputs": [],
       "source": [
         "image = preprocess(\"https://github.com/pytorch/hub/raw/master/images/dog.jpg\")\n",
-        "results = tr.server_predict({\"inputs\": image})\n",
+        "results = tr.predict({\"inputs\": image})\n",
         "postprocess(results[\"predictions\"])"
       ]
     }

docs/notebooks/xgboost_example.ipynb

@@ -63,9 +63,9 @@
       },
       "outputs": [],
       "source": [
-        "from truss import mk_truss\n",
+        "from truss import create\n",
         "\n",
-        "tr = mk_truss(model, target_directory=\"xgboost_truss\")"
+        "tr = create(model, target_directory=\"xgboost_truss\")"
       ]
     },
     {
@@ -76,7 +76,7 @@
       },
       "outputs": [],
       "source": [
-        "tr.server_predict({\"inputs\": [[0, 0, 0, 0, 0, 0]]})"
+        "tr.predict({\"inputs\": [[0, 0, 0, 0, 0, 0]]})"
       ]
     }
   ],

docs/reference/cli.md

@@ -79,10 +79,10 @@ Invokes the packaged model, either locally or in a Docker container.
 * `target_directory`: A Truss directory. If none, use current directory.
 * `request`: String formatted as json that represents request
 * `build_dir`: Directory where context is built. If none, a temp directory is created.
-* `tag`: Docker build image tag
-* `port`: Local port used to run image
-* `run_local`: Flag to run prediction locally (instead of on Docker)
-* `request_file`: Path to json file containing the request
+* `tag`: Docker build image tag.
+* `port`: Local port used to run image.
+* `use_docker`: Flag to run prediction on Docker, defaults to `False`.
+* `request_file`: Path to json file containing the request.
 
 ### `run-example`