HalfBlank-NewLesson1-DigitalImages.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Lesson 1: Digital Images\n",
    "\n",
    "Our goal today is to understand what a digital image is and how it is commonly represented as bits and bytes. We will cover the following topics:\n",
    "\n",
    "- 1. Loading and viewing images\n",
    "   - 1-1 Load image into iPython notebook   \n",
    "   - 1-2 Load the associated metadata   \n",
    "   - 1-3 Get bit depth of image    \n",
    "   - 1-4 View image in iPython notebook\n",
    "- 2. Using JSON and python dictionaries for metadata\n",
    "- 3. Indexing and arrays\n",
    "    - 2-1 Print a subset of pixel values (such as the top right corner)   \n",
    "    - 2-2 Set that subset of values to 0   \n",
    "    - 2-3 View the modified image in iPython       \n",
    "- 4. Histogram the pixel values in the image\n",
    "- 5. Color maps and color science\n",
    "- 6. File size, disk space, and memory: \n",
    "    - 5-1 Report the size of the file on the disk    \n",
    "    - 5-2 Report the size of the file read into memory       \n",
    "- 7. Errors and debugs:   \n",
    "    - 6-1 SyntaxError\n",
    "    - 6-2 NameError\n",
    "    - 6-3 IndexError   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 1** Practice using & navigating jupyter notebooks\n",
    "\n",
    "In the cell below (\"My notes about jupyter notebooks\"), change the cell type to markdown (comments and notes). It should look like this line of text when you're done. \n",
    "\n",
    "How? \n",
    "Click to the left of the cell. The marker on the right should turn blue (and not be green). If green, hit escape. \n",
    "\n",
    "Hint: You can find the shortcut for this by clicking outside of the text boxes and pressing \"h\". This information will be in the \"Command Mode\" section of the help menu."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "**My notes about jupyter notebooks**\n",
    "\n",
    "From Edit mode (green-selected cell that you can type into)\n",
    "   - SHIFT-ENTER runs the cell and selects the cell below\n",
    "   - ESC puts you into command mode\n",
    "   - h pulls up a help window\n",
    "\n",
    "From Command mode (blue-marked cell)\n",
    "   - a enters a new cell above the selected cell\n",
    "   - b enters a new cell below the selected cell\n",
    "   - dd deletes the selected cell\n",
    "   - y changes the cell type to python code\n",
    "   - m changes the cell to markdown\n",
    "   - ENTER enters edit mode"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.Loading and viewing images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First some boilerplate code to make it easier to access useful libraries, and to make it easier to visualize data in the notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline                \n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt   \n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\"%matplotlib inline\" is a magic function to sets the backend of matplotlib to the 'inline' backend, which means plotting right after commond, and saving the results.\n",
    "\n",
    "Now the `numpy` numerical array library is available as `np`. Plotting functions are available with `plt`, and `seaborn`'s advanced plots are accessed through `sns`. Just importing `seaborn` at all makes `Matplotlib` look nicer."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Let's set some defaults for the packages we just imported, which gets rid of grid line on our image plots!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use bash commands to:\n",
    "1. Create a folder on your desktop for this minicourse, call it datalucence2018\n",
    "2. Create two subfolders, one called code and a second called data\n",
    "    ~/Desktop/datalucence2018/code\n",
    "    ~/Desktop/datalucence2018/data\n",
    "2. Download the Data_DrugConfocalPanel folder from Canvas (under files)\n",
    "3. Move this folder into ~/Desktop/datalucence2018/data\n",
    "4. Navigate to datalucence2018\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/Tomato/Desktop/datalucence2018/data\n"
     ]
    }
   ],
   "source": [
    "cd ~/Desktop/datalucence2018/data/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/Users/Tomato/Desktop/datalucence2018/data'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[34mData_ConfocalDrugPanel\u001b[m\u001b[m/\r\n"
     ]
    }
   ],
   "source": [
    "ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The images in this folder are part of an experiment to determine the effects of drug A on the cells in the image. DMSO is the solvent/vehicle for drugA.  Thus the DMSO.tif file is the control condition and the drugA.tif is the test condition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[31mDMSO.tif\u001b[m\u001b[m*            \u001b[31mdrugA.tif\u001b[m\u001b[m*           \u001b[31mtemp_DMSO.tif\u001b[m\u001b[m*\r\n",
      "\u001b[31mDMSO_metadata.json\u001b[m\u001b[m*  \u001b[31mdrugA_metadata.json\u001b[m\u001b[m*\r\n"
     ]
    }
   ],
   "source": [
    "ls ./Data_ConfocalDrugPanel/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's make a common data path for our class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_path = \"../data/Data_ConfocalDrugPanel/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1-1 Load image into Jupyter notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from skimage.io import imread"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "imread?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "ctl_data = imread(data_path + \"DMSO.tif\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9, 819, 819, 3)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`data` is a `numpy` array. it's shape indicates that it has four dimensions.  So, this file is more complex than a single two-dimensional data array.  Fortunately, our former lab member (or former self) has left behind a metadata file to decode what data are in this file."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the metadata:\n",
    "Our former labmate has put the metadata into a file format called JSON. JSON files are easily loaded into python as the dictionary data type. Dictionaries in python are indexed with keys, which are strings instead of numerical indices (such as used in lists). To understand this concept, load the JSON file and examine it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#open JSON file in READ ONLY mode and put its contents into a dictionary\n",
    "import json #import JSON library\n",
    "with open(data_path + \"DMSO_metadata.json\", mode='r') as metadata_ctl:\n",
    "    meta_ctl = json.load(metadata_ctl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'cell_type': 'fibroblast', 'axes': ['ZXYTC'], 'pixel_size': [0.3459442, 0.3459442, 0.7694383], 'raw_data_date': ['add_this'], 'image_preprocessing_done': ['Noise_generator.ipynb', 'Branch=day3-day4-rework', \"GitCommitId=b'8669e48fb072e2f87f655b3e87f09e4926dbf1b0\\\\n'\"], 'channels': ['your_fav_protein', 'nucleus', 'actin']}\n"
     ]
    }
   ],
   "source": [
    "print(meta_ctl)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is a bit difficult to digest, so let's work our way through the information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cell_type\n",
      "axes\n",
      "pixel_size\n",
      "raw_data_date\n",
      "image_preprocessing_done\n",
      "channels\n"
     ]
    }
   ],
   "source": [
    "#list the keys in the dictionary\n",
    "for key in meta_ctl:\n",
    "    print(key)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cell_type fibroblast\n",
      "axes ['ZXYTC']\n",
      "pixel_size [0.3459442, 0.3459442, 0.7694383]\n",
      "raw_data_date ['add_this']\n",
      "image_preprocessing_done ['Noise_generator.ipynb', 'Branch=day3-day4-rework', \"GitCommitId=b'8669e48fb072e2f87f655b3e87f09e4926dbf1b0\\\\n'\"]\n",
      "channels ['your_fav_protein', 'nucleus', 'actin']\n"
     ]
    }
   ],
   "source": [
    "for key, value in meta_ctl.items():\n",
    "    print(key, value)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['your_fav_protein', 'nucleus', 'actin']"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "meta_ctl[\"channels\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 2: Look at the information imported from the metadata file\n",
    "    print out the axes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['ZXYTC']"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "meta_ctl[\"axes\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It can be useful to organize your data into a dict instead of a numerical array when one of the dimensions of the array corresponds to something that is non-numerical in nature. Here, the channel dimension is stored as another dimension in the numerical array that is wt_data. To get the image corresponding to one of the channels, you would have to remember which of the channel slices corresponds to the channel you would like to see. Below we'll organize the data into a dict so that the channels can be indexed by an intuitive string and not a numerical index."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9, 819, 819, 3)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Challenge: \n",
    "Based on the metadata and the shape, what is a likely meaning of the four axes of ctl_data?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2\n"
     ]
    }
   ],
   "source": [
    "#use the metadata to create a new dictionary that refers to one slice and all channeles\n",
    "ctl_slice = {}\n",
    "\n",
    "for idx, channel in enumerate(meta_ctl['channels']):\n",
    "    ctl_slice[channel] = ctl_data[3,:,:,idx]\n",
    "    print(idx)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### 1-3 Get bit depth of image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('uint16')"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_data.dtype"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "They are many kinds of data types. Here the data type is `uint16`.\n",
    " `uint16` means \"unsigned (not negative) integer with 16 bits per pixel\".\n",
    "\n",
    "16 bits means there are $2^{16} = 65536$ possible pixel values. Many scientific images will use 8 or even 12 bits which will have less contrast."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1-4 View image in iPython notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set_style(\"dark\") #use seaborn package to suppress grid lines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(819, 819)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"].shape #get the dimensions of this single slice in pixels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1c141812b0>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(ctl_slice[\"actin\"], cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We had to specify how we wanted our colorless image to be rendered on our colorful screen, which is why we included `cmap=gray` (cmap for colormap).\n",
    "\n",
    "We will talk more about colormaps and channels later."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Indexing and arrays"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-1 Print a subset of pixel values "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How would we index into the upper left-most pixel?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1039"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][0,0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What about the lower left?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "950"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][-1,0] # row -1 aka last, column 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What about a 10x10 slice from the upper right?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1123, 1199, 1067, 1069, 1096, 1108, 1069, 1052,  949],\n",
       "       [1187, 1154, 1109, 1052, 1032, 1059, 1085, 1101, 1039],\n",
       "       [1117, 1114, 1180, 1150, 1117, 1129, 1134, 1027,  978],\n",
       "       [1072, 1122, 1321, 1194, 1039, 1103, 1148, 1065, 1050],\n",
       "       [1106, 1154, 1144, 1087, 1107, 1121, 1131, 1011, 1065],\n",
       "       [1210, 1101, 1004, 1046, 1061, 1103, 1033, 1121, 1120],\n",
       "       [1204, 1153, 1085, 1143, 1096, 1108, 1215, 1164, 1131],\n",
       "       [1178, 1159, 1220, 1108, 1107, 1147, 1104, 1102, 1117],\n",
       "       [1096, 1176, 1137,  972, 1174, 1086, 1182, 1093, 1065],\n",
       "       [1117, 1139, 1183, 1134, 1023, 1120, 1097, 1141,    0]],\n",
       "      dtype=uint16)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][0:10,-10:-1] # row from 0 to 10 (not inclusive for 10! and column 10th-to-last second)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Does this look like the upper right of the source image?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1c17a44cc0>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(ctl_slice[\"actin\"][0:10,-10:-1], cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We only have 90 pixels not 100 pixels.\n",
    "\n",
    "Note that ranges of indices are exclusive on the high side, inclusive on the low. What happens if I have a slice `1:2`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[977]], dtype=uint16)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][1:2, 1:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We get row and column \"from 1, up to (but not including) 2\". This is the same as `data[1,1]`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "How can we get to the last column?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1123, 1199, 1067, 1069, 1096, 1108, 1069, 1052,  949, 1097],\n",
       "       [1187, 1154, 1109, 1052, 1032, 1059, 1085, 1101, 1039, 1018],\n",
       "       [1117, 1114, 1180, 1150, 1117, 1129, 1134, 1027,  978, 1095],\n",
       "       [1072, 1122, 1321, 1194, 1039, 1103, 1148, 1065, 1050, 1024],\n",
       "       [1106, 1154, 1144, 1087, 1107, 1121, 1131, 1011, 1065, 1074],\n",
       "       [1210, 1101, 1004, 1046, 1061, 1103, 1033, 1121, 1120, 1100],\n",
       "       [1204, 1153, 1085, 1143, 1096, 1108, 1215, 1164, 1131, 1120],\n",
       "       [1178, 1159, 1220, 1108, 1107, 1147, 1104, 1102, 1117, 1084],\n",
       "       [1096, 1176, 1137,  972, 1174, 1086, 1182, 1093, 1065, 1018],\n",
       "       [1117, 1139, 1183, 1134, 1023, 1120, 1097, 1141,    0, 1060]],\n",
       "      dtype=uint16)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][0:10, 809:819]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To save some typing when slicing into your data, we can leave off the value before the colon, meaning (to the beginning). Leaving off the value after means (to the end)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1123, 1199, 1067, 1069, 1096, 1108, 1069, 1052,  949, 1097],\n",
       "       [1187, 1154, 1109, 1052, 1032, 1059, 1085, 1101, 1039, 1018],\n",
       "       [1117, 1114, 1180, 1150, 1117, 1129, 1134, 1027,  978, 1095],\n",
       "       [1072, 1122, 1321, 1194, 1039, 1103, 1148, 1065, 1050, 1024],\n",
       "       [1106, 1154, 1144, 1087, 1107, 1121, 1131, 1011, 1065, 1074],\n",
       "       [1210, 1101, 1004, 1046, 1061, 1103, 1033, 1121, 1120, 1100],\n",
       "       [1204, 1153, 1085, 1143, 1096, 1108, 1215, 1164, 1131, 1120],\n",
       "       [1178, 1159, 1220, 1108, 1107, 1147, 1104, 1102, 1117, 1084],\n",
       "       [1096, 1176, 1137,  972, 1174, 1086, 1182, 1093, 1065, 1018],\n",
       "       [1117, 1139, 1183, 1134, 1023, 1120, 1097, 1141,    0, 1060]],\n",
       "      dtype=uint16)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ctl_slice[\"actin\"][:10,-10:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-2 Set that subset of values to 0 "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's not ruin our original by making a copy! (Keep RAW DATA RAW)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "modified_data = ctl_slice[\"actin\"].copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We've viewed data using slicing, now let's set data using slicing!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "modified_data[:500,-500:] = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Even though `modified_data[:500,:500]` is a 500x500 array, and 0 is just a scalar, `numpy` is smart and will _broadcast_ the 1x1 value `0` so that the whole 500x500 array is set to a 500x500 array of zeros."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2-3 View the modified image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1c17af6470>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(modified_data, cmap='gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Histogram the pixel values in the image "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pixels in an image are just represented by numbers. We can get a sense for the distribution of brightness in our image by looking at a histogram of intensities. Here we don't think about an image as representing something spacial - just a collection of numbers.\n",
    "\n",
    "Make our array into a simple 1D list of data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "flat_data = ctl_slice[\"nucleus\"].flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.distplot?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/Tomato/anaconda3/lib/python3.5/site-packages/matplotlib/axes/_axes.py:6462: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(0, 20000)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(flat_data, norm_hist=True) #distplot=distribution plot\n",
    "plt.xlabel(\"Pixel intensity\")\n",
    "plt.ylabel(\"Frequency in the image\")\n",
    "plt.xlim(0,20000)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4.Color maps and color science"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we wanted to visualize things in a more striking way, with false colors and more contrast, we could use a different colormap."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Jet is a bad colormap because it is perceptually non-uniform."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Access IPython's browser-based image display."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quickly display a diagram we saved in a file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The jumps in the \"perceptual deltas\" plot are values where this colormap makes it look like sharp transitions occur when they do not. Colorblind users rely more on lightness than hue, and will be especially misled."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This colormap is perceptually uniform by design! It looks cool too."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Exercise 3** We discussed that \"jet\" or really any of the rainbow colormaps are perceptually deceiving. Find the documentation for the matplotlib colormaps online. View your \"data\" slice with your perceptually-uniform colormap of choice. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here is the link:\n",
    "https://matplotlib.org/users/colormaps.html"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "View image using plasma colormap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following cell to look up the imshow documentation on how to adjust the contrast by specifying vmin and vmax. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. File size, disk space, and memory "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5-1 Report the size of the file on the disk "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "a human-readable description of the image file we've been using\n",
    "Pro tip: a leading `!` in a notebook drops you into bash."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note the \"35M\". That's our file size."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5-2 Report the size of the file read into memory "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What is this in MB?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So our image on-disk and loaded into Python are the same size. Therefore, this was an _uncompressed_ or _raw_ tif. Such files are quick to read and write, but take up lots of space on your hard drive. Image files always take up $bitdepth \\times x \\times y \\times z$ in memory."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Errors and debugs "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_image(filename):\n",
    "    return np.random.rand(400, 600)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6-1 Syntax errors"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What is this code supposed to do?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50)\n",
    "    image_data = load_image(\"image_{}\".format(i))\n",
    "    my_images.append(image_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This file is trying to load a series of numbered images into an array of images.\n",
    "\n",
    "_However_ running it gives you an error\n",
    "\n",
    "Python errors are given to you as a \"stack trace\". Let's first dissect a stack trace.\n",
    "\n",
    "**`File \"<ipython-input-some_number-some_ID_number>\", line 3`** tells you what file, and where in the file, the error came from. Because we're in a notebook, instead of a file we get a message that tells us we're in a notebook using IPython and that the error was in line 3.\n",
    "\n",
    "**`  for i in range(50)`** is where Python conveniently reminds us what code was at line 3. Sometimes it shows us a bit of code before and after to give us context.\n",
    "\n",
    "**`^`** marks exactly _where_ in line 3 the problem was noticed. This can be tricky because a problem with a function call may not be noticed until the closing ')' at the very end of the function. Nonetheless, here it might be helpful.\n",
    "\n",
    "**`SyntaxError: invalid syntax`**. This tells us that our problem is a SyntaxError, which is one of a large hierarchy of errors python can provide us with.\n",
    "\n",
    "`SyntaxError`s happen when Python sees you violating the rules of the language - it's a problem with the literal code characters you have typed rather than what your code is trying to do conceptually. Therefore they are usually short errors with quick fixes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Do you see how to fix _this_ error?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50)\n",
    "    image_data = load_image(\"image_{}\".format(i))\n",
    "    my_images.append(image_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.2 NameError "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`NameError`s happen when you try to access a python variable or function that does not exist yet. To get a NameError, Python has to actually try to run your code, so NameError is a type of `RuntimeError`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Can you add in a line that fixes this code?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50):\n",
    "    image_data = load_image(\"image_{}\".format(i))\n",
    "    my_images.append(image_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case \"my_images\" does not exist, so it cannot be appended to, thus the name error. Making an empty my_images array solves this problem.\n",
    "\n",
    "Now that our images are loaded, let's check to make sure each images's mean intensity changes by less than 10% compared with the next timage. This might be a good check to make sure no one bumped our microscope or turned on a light while we were taking images."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50):\n",
    "    intensity = my_images[i].mean()\n",
    "    next_intensity = my_images[i+1].mean()\n",
    "    if abs(intensity - next_intensity) > 0.10 * intensity:\n",
    "        print(\"Notice: intensity jumped between images {} and {}\".format(i, i+1))\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.3 IndexError"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`IndexError`s happen when you try to access data from a list-like object, such as a `numpy` array or image, but the location you requested does not exist. Like asking for index 10 for a list of 10 items.\n",
    "\n",
    "In image processing, this is often caused by switching your rows/columns or width/height. Say you have a 400x600 image and try to access a pixel at row 401-600.\n",
    "\n",
    "It is also generally common when looping through something by index. This is a good reason to use Python's `for item in collection` syntax rather than looping through indices!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = [1, 3, \"cat\", 0.4]\n",
    "\n",
    "# This is less clear and prone to error\n",
    "for index in range(4):\n",
    "    item = data[index]\n",
    "    print(item)\n",
    "\n",
    "# Than this\n",
    "for item in data:\n",
    "    print(item)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One interesting thing about Python loop's current value, `i` in this case, is available outside the loop and does not reset until you run the loop again. This let's us quickly check what value `i` took on when the code crashed!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now can you explain what happened here? Can you fix this code?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50):\n",
    "    intensity = my_images[i].mean()\n",
    "    next_intensity = my_images[i+1].mean()\n",
    "    if abs(intensity - next_intensity) > 0.10 * intensity:\n",
    "        print(\"Notice: intensity jumped between images {} and {}\".format(i, i+1))\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sufficiency and necessity of error messages"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What would have happened if instead of comparing image `i` to image `i+1` we had compared image `i-1` (previous image) to `i`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(50):\n",
    "    intensity = my_images[i-1].mean()\n",
    "    next_intensity = my_images[i].mean()\n",
    "    if abs(intensity - next_intensity) > 0.10 * intensity:\n",
    "        print(\"Notice: intensity jumped between images {} and {}\".format(i-1, i))\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exercise There was no error! Did this solve our problem?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No, it did not. Consider the first pass through the loop. `i` is 0. `i-1` is -1. What is `my_images[-1]`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`my_images[-1]` doesn't give a name error, -1 refers to the _last_ item in the list. But we don't want to compare the first and last images, so although this runs it's not the right behavior."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Long stack traces"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What can make the errors that we see in this course particularly daunting is that we use many libraries which use other libraries which in turn use more libraries, etc. This means that the peice of code that reports the error is often code we didn't write or didn't even know was being run, which can make errors feel unfair or unsolveable. But here's what's really happening.\n",
    "\n",
    "Imagine that you send me on an errand to buy groceries. You give me detailed instructions (a program) describing the steps to take. So I get in the car and start driving to the store. Half way there, I notice I'm out of gas. That's OK, I have my own program for dealing with that. \"Buy gas at a station\" is a bit of instructions you didn't know I had, nor did you anticipate me using, but it's being used now anyway. I pull into the gas station, get out of my car, pay for the gas, and try to start filling up. However, the gas cover locks from the inside. I don't know about this (new car), and send you a text trying to precisely describe the error: \"Nozzle cannot pass through solid metal\".\n",
    "\n",
    "So here you are, having sent me to buy groceries, and I tell you I can't because \"nozzle cannot pass through solid metal\". Python is frustrating in the same way: generally it tells you the _lowest-level problem_ when it fails. It's important to keep this in mind: the piece of code that reports an error is probably not the one that caused it!\n",
    "\n",
    "Stack traces report errors with the first call at the top. **This means that you should read a long stack trace by starting at the bottom and working up until you see a line of code that you wrote or a function you called.** This line is likely to be the line you have to change. Maybe you passed a string to a function when it needed an integer.\n",
    "\n",
    "Lines below code you wrote in the stack trace may contain hints about what's wrong. Maybe the lowest error at the bottom of the stack trace is \"Cannot subtract string from int\", which is a cluse that there was a string where an int should be.\n",
    "\n",
    "Higher lines tell you the context that the error happened in, i.e. what code ran before the error. Maybe you have a function that you use several times - you want to know _which_ usage of the function is giving you the error. If it's the 2nd time you use the function, then either the function _can_ work, but not in a particular context, or the function is being called in the 2nd location _before_ the first: the flow of your program is not what you expected."
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}