development.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Have trained a classifier following https://blog.machinebox.io/how-anyone-can-build-a-machine-learning-image-classifier-from-photos-on-your-hard-drive-very-5c20c6f2764f using my own bird data. Achieved an accuracy of 88% - but the first time I ran the script (different random selection of images) accuracy was about 92%, so clearly I am very sensitive to the images randomly selected and should add more images:\n",
    "```\n",
    "Correct:    180\n",
    "Incorrect:  23\n",
    "Errors:     0\n",
    "Accuracy:   88.66995073891626%\n",
    "```\n",
    "\n",
    "https://machineboxio.com/docs/classificationbox\n",
    "\n",
    "classificationbox is online learning (supervised learning), it works also with little data but it build the classifier function so it needs more data. for my bird monirtoring project my 2 classes would be bird/no_bird, and I just post the image data\n",
    "\n",
    "Classifiers can be made to help solve a wide range of example use cases, for example:\n",
    "\n",
    "* Learn about how your company is perceived by grouping tweets into positive and negative\n",
    "* Automatically group photos of cats and dogs\n",
    "* Group emails into spam and non-spam categories\n",
    "* Build a classifier to detect the language of a piece of text based on previously taught examples\n",
    "\n",
    "```\n",
    "sudo docker pull machinebox/classificationbox\n",
    "\n",
    "sudo docker run -p 8080:8080 -e \"MB_KEY=$MB_KEY\" machinebox/classificationbox\n",
    "```\n",
    "http://localhost:8080/\n",
    "\n",
    "Note that if you restart classificationbox models will be lost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "# import curlify\n",
    "import re\n",
    "import json\n",
    "import base64\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as patches\n",
    "%matplotlib inline\n",
    "\n",
    "def base64_encode_file(file_path):\n",
    "    \"\"\"\n",
    "    Takes the path to an image and returns the base64 encoded\n",
    "    image data as a string.\n",
    "    \"\"\"\n",
    "    with open(file_path, \"rb\") as f:\n",
    "        file_data = base64.b64encode(f.read()).decode('ascii')\n",
    "    return file_data\n",
    "\n",
    "ATTR_ID = 'id'\n",
    "ATTR_CONFIDENCE = 'confidence'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "IP =  'localhost'\n",
    "#IP =  '192.168.0.30'\n",
    "PORT = '8080'\n",
    "CLASSIFIER = 'classificationbox'\n",
    "CONFIDENCE = 80\n",
    "\n",
    "MODELS_URL = 'http://{}:{}/{}/models'.format(IP, PORT, CLASSIFIER)\n",
    "MODEL_CREATION_URL = 'http://{}:{}/{}/models'.format(IP, PORT, CLASSIFIER)\n",
    "STATE_POST_URL = 'http://{}:{}/classificationbox/state'.format(IP, PORT)\n",
    "\n",
    "username = 'my_username'\n",
    "password = 'my_password'\n",
    "kwargs = {}\n",
    "if username:\n",
    "    kwargs['auth'] = requests.auth.HTTPBasicAuth(username, password)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'http://localhost:8080/classificationbox/models'"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODELS_URL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_models(url):\n",
    "    \"\"\"Return the list of models.\"\"\"\n",
    "    try:\n",
    "        response = requests.get(url, **kwargs)\n",
    "        response_json = response.json()\n",
    "        if response_json['success']:\n",
    "            if len(response_json['models']) == 0:\n",
    "                print(\"%s error: No models found\", CLASSIFIER)\n",
    "            else:\n",
    "                return response_json['models']\n",
    "    except requests.exceptions.ConnectionError:\n",
    "        print(\"ConnectionError: Is %s running?\", CLASSIFIER)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'auth': <requests.auth.HTTPBasicAuth at 0x109e79a90>}"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "kwargs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'success': True,\n",
       " 'models': [{'id': '5b0ce5d8023d4e35', 'name': '5b0ce5d8023d4e35'}]}"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "response = requests.get(MODELS_URL, timeout=9, **kwargs)\n",
    "response_json = response.json()\n",
    "response_json"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets list my models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'id': '5b0ce5d8023d4e35', 'name': '5b0ce5d8023d4e35'}]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models = get_models(MODELS_URL)\n",
    "models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5b0ce5d8023d4e35 5b0ce5d8023d4e35\n"
     ]
    }
   ],
   "source": [
    "for model in models:\n",
    "    print(model['id'], model['name'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I want the model which was generated by the GO script on my bird data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'5b0ce5d8023d4e35'"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_ID = models[0]['id']\n",
    "MODEL_ID"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL_PREDICT_URL = 'http://{}:{}/{}/models/{}/predict'.format(IP, PORT, CLASSIFIER, MODEL_ID)\n",
    "MODEL_STATS_URL = 'http://{}:{}/{}/models/{}/stats'.format(IP, PORT, CLASSIFIER, MODEL_ID)\n",
    "\n",
    "MODEL_STATE_URL = 'http://{}:{}/classificationbox/state/{}'.format(IP, PORT, MODEL_ID)\n",
    "MODEL_TEACH_URL = 'http://{}:{}/{}/models/{}/teach'.format(IP, PORT, CLASSIFIER, MODEL_ID)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now lets see the stats on this model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'success': True, 'predictions': 1, 'examples': 0, 'classes': []}"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_stats = requests.get(MODEL_STATS_URL).json()\n",
    "model_stats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lets make a prediction on a bird image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10c96eef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "IMG_FILE = \"bird_project/bird.jpg\"\n",
    "FIG_SIZE = (6, 4)\n",
    "\n",
    "img = plt.imread(IMG_FILE)\n",
    "fig, ax = plt.subplots(figsize=FIG_SIZE)\n",
    "ax.imshow(img);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'http://localhost:8080/classificationbox/models/5b0ce5d8023d4e35/predict'"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_PREDICT_URL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predict via JSON encoded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "predict_data  = {\n",
    "    \"inputs\": [\n",
    "        {\"key\": \"image\", \"type\": \"image_base64\", \"value\": base64_encode_file(IMG_FILE)}]}\n",
    "\n",
    "try:\n",
    "    response = requests.post(MODEL_PREDICT_URL, json=predict_data)\n",
    "except ValueError:\n",
    "    print(\"Classificationbox error: {}\".format(response.json.text))\n",
    " #   response = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(curlify.to_curl(response.request))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'success': True,\n",
       " 'classes': [{'id': 'birds', 'score': 0.915892},\n",
       "  {'id': 'not_birds', 'score': 0.084108}]}"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "requests_json = response.json()\n",
    "requests_json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_classes(api_classes):\n",
    "    \"\"\"Parse the API classes data into the format required.\"\"\"\n",
    "    parsed_classes = []\n",
    "    for entry in api_classes:\n",
    "        class_ = {}\n",
    "        class_[ATTR_ID] = entry['id']\n",
    "        class_[ATTR_CONFIDENCE] = round(entry['score'] * 100.0, 2)\n",
    "        parsed_classes.append(class_)\n",
    "    return parsed_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'id': 'birds', 'confidence': 91.59}, {'id': 'not_birds', 'confidence': 8.41}]"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes = parse_classes(requests_json['classes'])\n",
    "classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'birds'"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes[0]['id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_classes(classes_json):\n",
    "    \"\"\"Return the classes data.\"\"\"\n",
    "    classes_dict = {class_result['id']: round(class_result['score'] * 100.0, 2)\n",
    "                    for class_result in classes_json}\n",
    "    return classes_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'birds': 91.59, 'not_birds': 8.41}"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes = get_classes(response.json()['classes'])\n",
    "classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "birds with confidence 91.59\n"
     ]
    }
   ],
   "source": [
    "for key, value in classes.items():\n",
    "    if value >= CONFIDENCE:\n",
    "        print(\"{} with confidence {}\".format(key, value))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download model\n",
    "We can download the model as a binary file. A `model_{model_id}.classificationbox` data file will be downloaded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL_STATE_URL # Pasting this in the browser will download the files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "response = requests.get(MODEL_STATE_URL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = \"model_{}.classificationbox\".format(MODEL_ID)\n",
    "filename"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "open(filename, 'wb').write(response.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Upload model\n",
    "Now lets upload the model having restarted connection box"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "http://localhost:8080/classificationbox/state/5b0ce5d8023d4e35\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'success': True,\n",
       " 'id': '5b0ce5d8023d4e35',\n",
       " 'name': '5b0ce5d8023d4e35',\n",
       " 'options': {},\n",
       " 'predict_only': False}"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "MODEL_ID = '5b0ce5d8023d4e35'\n",
    "MODEL_STATE_URL = 'http://{}:{}/classificationbox/state/{}'.format(IP, PORT, MODEL_ID)\n",
    "print(MODEL_STATE_URL)\n",
    "filename = '/Users/robincole/Documents/Data/Machinebox/classificationbox/birds.classificationbox'\n",
    "model_data  = {\"base64\": base64_encode_file(filename)}\n",
    "\n",
    "requests.post(STATE_POST_URL, json=model_data, **kwargs).json()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Class\n",
    "Make a class wrapper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "CLASSIFIER = 'classificationbox'\n",
    "\n",
    "class ClassificationboxEntity():\n",
    "    \"\"\"Perform an image classification.\"\"\"\n",
    "\n",
    "    def __init__(self, ip, port, camera_entity, model_id, model_name):\n",
    "        \"\"\"Initialise a classificationbox model entity.\"\"\"\n",
    "        self._base_url = \"http://{}:{}/{}/\".format(ip, port, CLASSIFIER)\n",
    "        self._camera = camera_entity\n",
    "\n",
    "        self._model_id = model_id\n",
    "        self._model_name = model_name\n",
    "\n",
    "        camera_name = camera_entity # split_entity_id(camera_entity)[1]\n",
    "        self._name = \"{} {} {}\".format(\n",
    "            CLASSIFIER, camera_name, model_name)\n",
    "        \n",
    "    @property\n",
    "    def name(self):\n",
    "        \"\"\"Return the name of the sensor.\"\"\"\n",
    "        return self._name\n",
    "    \n",
    "    @property\n",
    "    def device_state_attributes(self):\n",
    "        \"\"\"Return the classifier attributes.\"\"\"\n",
    "        return {\n",
    "            'model_id': self._model_id,\n",
    "            'model_name': self._model_name\n",
    "            }\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "entities = []\n",
    "\n",
    "if models_query['success']:\n",
    "    for model in models_query['models']:\n",
    "        camera_entity = 'mock_cam'\n",
    "        entity = ClassificationboxEntity(IP, PORT, camera_entity, model['id'], model['name'])\n",
    "        entities.append(entity)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "entities[1].name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "entities[1].device_state_attributes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    response = requests.get(MODELS_LIST_URL, timeout=9).json()\n",
    "except requests.exceptions.ConnectionError:\n",
    "    _LOGGER.error(\"ConnectionError: Is %s running?\", CLASSIFIER)\n",
    "\n",
    "if response.staus_code == 200:\n",
    "    print('success')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "requests.get(MODELS_LIST_URL, timeout=9).status_code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODELS_LIST_URL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}