TensorPy · mdmintz · Oct 20, 2016 · Oct 20, 2016 · Oct 20, 2016 · Oct 20, 2016
diff --git a/Dockerfile b/Dockerfile
@@ -37,23 +37,19 @@ RUN locale-gen ${LANGUAGE} \
 #==============================
 # Set up TensorFlow / TensorPy
 #==============================
-COPY docker/docker_install.sh /TensorPy/docker_install.sh
-COPY download_imagenet.py /TensorPy/download_imagenet.py
-COPY classify.py /TensorPy/classify.py
+COPY third_party/docker/docker_install.sh /TensorPy/docker_install.sh
 COPY requirements.txt /TensorPy/
 COPY setup.py /TensorPy/
 COPY tensorpy /TensorPy/tensorpy/
 COPY examples /TensorPy/examples/
-COPY sample_images /TensorPy/sample_images/
-COPY test_classify_image.sh /TensorPy/
-COPY docker /TensorPy/docker/
-COPY docker/run_docker_test.sh /TensorPy/
-RUN cd /TensorPy && ls && ./docker/docker_install.sh
+COPY third_party/docker /TensorPy/third_party/docker/
+COPY third_party/docker/run_docker_test.sh /TensorPy/
+RUN cd /TensorPy && ls && ./third_party/docker/docker_install.sh
 RUN cd /TensorPy && pip install -r requirements.txt
 
 #===================
 # Create entrypoint 
 #===================
-COPY docker/docker-entrypoint.sh /
+COPY third_party/docker/docker-entrypoint.sh /
 ENTRYPOINT ["/docker-entrypoint.sh"]
 CMD ["/bin/bash"]
diff --git a/README.md b/README.md
@@ -3,14 +3,18 @@
 
 [![pypi](https://img.shields.io/pypi/v/tensorpy.svg)](https://pypi.python.org/pypi/tensorpy) [![Python version](https://img.shields.io/badge/python-2.7-22AADD.svg "Python version")](https://docs.python.org/2/) [![MIT License](http://img.shields.io/badge/license-MIT-22BBCC.svg "MIT License")](https://github.com/TensorPy/TensorPy/blob/master/LICENSE) [![Join the chat at https://gitter.im/TensorPy/Lobby](https://badges.gitter.im/TensorPy/TensorPy.svg)](https://gitter.im/TensorPy/Lobby?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
 
-**Making TensorFlow easier to use with Python**
+**Image Classification with TensorFlow made easy!**
 
-You can use TensorPy to classify images by simply passing a URL. [TensorFlow](https://www.tensorflow.org/) image recognition does the hard work.
+You can use TensorPy to classify images by simply passing a URL on the command line, or by using TensorPy in your Python program. **[TensorFlow](https://www.tensorflow.org/)** does all the image-recognition work.
 
+TensorPy also simplifies TensorFlow installation by automating several setup steps into a single script (See **[install.sh](https://github.com/TensorPy/TensorPy/blob/master/install.sh)** for details).
 
-### Setup Instructions for Mac and Ubuntu/Linux
+See **[help_docs/how_tensorpy_works.md](https://github.com/TensorPy/TensorPy/blob/master/help_docs/how_tensorpy_works.md)** for a detailed explanation of how TensorPy works.
 
-(**Windows & Docker users**: See the [Docker ReadMe](https://github.com/TensorPy/TensorPy/blob/master/docker/ReadMe.md) for running on a Docker machine. Windows requires Docker to run TensorFlow.)
+
+## Setup Steps for Mac & Ubuntu/Linux
+
+(**Windows & Docker users**: See the **[Docker ReadMe](https://github.com/TensorPy/TensorPy/blob/master/third_party/docker/ReadMe.md)** for running on a Docker machine. Windows requires Docker to run TensorFlow.)
 
 #### **Step 1:** Create and activate a virtual environment named "tensorpy"
 
@@ -23,9 +27,9 @@ git clone https://github.com/TensorPy/TensorPy.git
 cd TensorPy
 ```
 
-#### **Step 3:** Install TensorPy, TensorFlow, and ImageNet
+#### **Step 3:** Install TensorPy, TensorFlow, and ImageNet/Inception
 
-One script installs them all. (See [install.sh](https://github.com/TensorPy/TensorPy/blob/master/install.sh) for details.)
+The **[install.sh](https://github.com/TensorPy/TensorPy/blob/master/install.sh)** script installs everything you need:
 
 ```bash
 ./install.sh
@@ -36,18 +40,24 @@ One script installs them all. (See [install.sh](https://github.com/TensorPy/Tens
 Classify a single image from a URL:
 
 ```bash
-classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/sample_images/happy_animal.jpg"
+classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg"
 ```
 
 Classify all images on a web page:
 
 ```bash
-classify "https://github.com/TensorPy/TensorPy/tree/master/sample_images"
+classify "https://github.com/TensorPy/TensorPy/tree/master/examples/images"
 ```
 
-Classify a single image from a URL from within a Python program: (See [test_python_classify.py](https://github.com/TensorPy/TensorPy/blob/master/examples/test_python_classify.py) for details.)
+Classify a single image from a URL from within a Python program: (See **[test_python_classify.py](https://github.com/TensorPy/TensorPy/blob/master/examples/test_python_classify.py)** for details.)
 
 ```bash
 cd examples
 python test_python_classify.py
 ```
+
+____________
+
+### Future Work:
+
+Eventually, the headline will change from "Image Classification with TensorFlow made easy!" to "Machine Learning with TensorFlow made easy!" once I expand on TensorPy to make other features of TensorFlow easier too. Stay tuned for updates!
diff --git a/docker/run_docker_test.sh b/docker/run_docker_test.sh
diff --git a/examples/images/ReadMe.md b/examples/images/ReadMe.md
@@ -0,0 +1,15 @@
+### Here are some cool animals!
+
+**(For testing image recognition)**
+
+File names are purposely made ambiguous to demonstrate that classifications are based on the image itself and not the image name. Classifications aren't always perfect: "water_animal.jpg" is an image of a dolphin, but TensorFlow classifies it as a type of shark. TensorFlow's neural net model can be trained to perform better (more on that later!).
+
+![](https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/bird_animal.jpg "Animal")
+
+![](https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/cat_animal.jpg "Animal")
+
+![](https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg "Animal")
+
+![](https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/reptile_animal.jpg "Animal")
+
+![](https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/water_animal.jpg "Animal")
diff --git a/sample_images/bird_animal.jpg → examples/images/bird_animal.jpg b/sample_images/bird_animal.jpg → examples/images/bird_animal.jpg
diff --git a/sample_images/cat_animal.jpg → examples/images/cat_animal.jpg b/sample_images/cat_animal.jpg → examples/images/cat_animal.jpg
diff --git a/sample_images/happy_animal.jpg → examples/images/happy_animal.jpg b/sample_images/happy_animal.jpg → examples/images/happy_animal.jpg
diff --git a/sample_images/reptile_animal.jpg → examples/images/reptile_animal.jpg b/sample_images/reptile_animal.jpg → examples/images/reptile_animal.jpg
diff --git a/sample_images/water_animal.jpg → examples/images/water_animal.jpg b/sample_images/water_animal.jpg → examples/images/water_animal.jpg
diff --git a/examples/test_classify_image.sh b/examples/test_classify_image.sh
@@ -1 +1 @@
-classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/sample_images/happy_animal.jpg"
+classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg"
diff --git a/examples/test_classify_page.sh b/examples/test_classify_page.sh
@@ -1 +1 @@
-classify "https://github.com/TensorPy/TensorPy/tree/master/sample_images"
+classify "https://github.com/TensorPy/TensorPy/tree/master/examples/images"
diff --git a/examples/test_python_classify.py b/examples/test_python_classify.py
@@ -1,5 +1,5 @@
 from tensorpy import image_base
 
 result = image_base.classify(
-    "https://raw.githubusercontent.com/TensorPy/TensorPy/master/sample_images/happy_animal.jpg")  # noqa
+    "https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg")  # noqa
 print("\nBest match classification:\n%s\n" % result)
diff --git a/help_docs/how_tensorpy_works.md b/help_docs/how_tensorpy_works.md
@@ -0,0 +1,7 @@
+### How TensorPy Works (Detailed Explanation)
+
+Once TensorPy is installed with [install.sh](https://github.com/TensorPy/TensorPy/blob/master/install.sh), a command called "classify" is added to your command line, which takes a URL as input. When called, TensorPy determines if that URL links to an image or a web image. If it's an image, TensorPy downloads the image to a new folder called "downloads_folder". From there, the image is converted to a JPEG. Then, the image is fed to the local [tensorpy/classify_image.py](https://github.com/TensorPy/TensorPy/blob/master/tensorpy/classify_image.py) which tells TensorFlow to use the ImageNet Inception database to classify the image and print out the result.
+
+If the "classify" command is called with a web page as input, then all images on that page (up to the limit defined in [tensorpy/settings.py](https://github.com/TensorPy/TensorPy/blob/master/tensorpy/settings.py)) will get downloaded to "downloads_folder", where TensorPy does the work listed above per image, with the exception of images smaller than the minimum size defined in tensorpy/settings.py (currently 50x50 pixels). Images that are too small will be skipped because results have shown that icons and other tiny images get extremely poor classification results.
+
+You can also use TensorPy to classify images from within a Python program. (See the example to learn how: [examples/test_python_classify.py](https://github.com/TensorPy/TensorPy/blob/master/examples/test_python_classify.py))
diff --git a/sample_images/ReadMe.md b/sample_images/ReadMe.md
diff --git a/setup.py b/setup.py
@@ -8,7 +8,7 @@
 
 setup(
     name='tensorpy',
-    version='1.0.5',
+    version='1.0.6',
     url='http://tensorpy.com',
     author='Michael Mintz',
     author_email='@mintzworld',

diff --git a/tensorpy/classify.py b/tensorpy/classify.py
@@ -61,7 +61,7 @@ def download_and_classify_image(image_url):
 
 
 def main():
-    expected_arg = "[PAGE_URL or IMAGE_URL]"
+    expected_arg = "[A valid PAGE_URL or IMAGE_URL]"
     num_args = len(sys.argv)
     if num_args < 2 or num_args > 2:
         print "\n* INVALID RUN COMMAND! *  Usage:"

diff --git a/third_party/ReadMe.md b/third_party/ReadMe.md
@@ -0,0 +1,3 @@
+## Third-Party / Integrations
+
+This folder contains TensorPy integrations
diff --git a/docker/ReadMe.md → third_party/docker/ReadMe.md b/docker/ReadMe.md → third_party/docker/ReadMe.md
diff --git a/docker/docker-entrypoint.sh → third_party/docker/docker-entrypoint.sh b/docker/docker-entrypoint.sh → third_party/docker/docker-entrypoint.sh
diff --git a/docker/docker_install.sh → third_party/docker/docker_install.sh b/docker/docker_install.sh → third_party/docker/docker_install.sh
@@ -8,5 +8,5 @@ echo "Initializing TensorFlow setup on an Ubuntu Docker machine..."
 export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.10.0-cp27-none-linux_x86_64.whl
 pip install $TF_BINARY_URL
 echo "Downloading ImageNet (image database for classifying images)..."
-python download_imagenet.py
+python tensorpy/download_imagenet.py
 exit
diff --git a/third_party/docker/run_docker_test.sh b/third_party/docker/run_docker_test.sh
@@ -0,0 +1,8 @@
+#!/bin/bash
+set -e
+# Run example test from inside Docker image
+cd /TensorPy/
+echo "Running example test: Classify a single image from a URL."
+echo "classify https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg"
+classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg"
+exec "$@"
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/sample_images/happy_animal.jpg"
		classify "https://raw.githubusercontent.com/TensorPy/TensorPy/master/examples/images/happy_animal.jpg"
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		classify "https://github.com/TensorPy/TensorPy/tree/master/sample_images"
		classify "https://github.com/TensorPy/TensorPy/tree/master/examples/images"
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,3 @@
		## Third-Party / Integrations

		This folder contains TensorPy integrations