reports/RQ4.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# RQ 4 (Encapsulation)\n",
    "\n",
    ">  Is unsafe code typically shielded from clients through safe abstractions?\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import itertools\n",
    "import functools\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import tree\n",
    "\n",
    "import utils\n",
    "import top_crates\n",
    "from whitelists import *\n",
    "from drawing import *\n",
    "\n",
    "# Configure pyplot: enlarge plots \n",
    "plt.rcParams['figure.figsize'] = [15, 8]\n",
    "# Configure pandas: show all columns when printing a dataframe\n",
    "pd.set_option('display.max_columns', None)\n",
    "# Configure pandas: show more rows when printing a dataframe\n",
    "pd.set_option('display.min_rows', 20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load tables\n",
    "function_definitions = utils.load_selected_function_definitions()\n",
    "selected_builds = utils.load_selected_builds()\n",
    "\n",
    "# Preprocessing\n",
    "unsafe_function_definitions = function_definitions[function_definitions.unsafety == \"Unsafe\"].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Validation\n",
    "assert function_definitions[function_definitions.unsafety == \"Unknown\"].abi.unique() == [\"Closure\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Whitelists\n",
    "all_function_names = function_definitions.def_path.unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Query 6\n",
    "\n",
    "> We then count *how many unsafe functions are (1) declared private, (2) visible within their crate,\n",
    "and (3) visible to other crates.*\n",
    "\n",
    "#### Results from the paper for comparison (Table 2)\n",
    "\n",
    "|Visibility| #functions      |    %  |\n",
    "| ------------- |--------------|-------|\n",
    "| Private       | 65,230       |  11.7 |\n",
    "| Restricted    | 1,535        |   0.3 |\n",
    "| Public        | 489,928      |  87.9 |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "unsafe_function_definitions: 557380\n",
      "The 'Unknown' are trait methods\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Public        489928\n",
       "Private        65230\n",
       "Crate           1378\n",
       "Unknown          687\n",
       "Restricted       157\n",
       "Name: visibility, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"unsafe_function_definitions:\", len(unsafe_function_definitions))\n",
    "\n",
    "# In our dataset the 'Unknown' rows coincide with the trait items\n",
    "assert (unsafe_function_definitions.visibility == \"Unknown\").equals(unsafe_function_definitions.is_trait_item)\n",
    "\n",
    "print(\"The 'Unknown' are trait methods\")\n",
    "unsafe_function_definitions.visibility.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of unsafe functions: 557380\n",
      "Total number of trait-method unsafe functions: 687 (0.123255%)\n",
      "Total number of non-trait-method unsafe functions:556693 (99.876745%)\n",
      "is_public        489928\n",
      "is_private        65230\n",
      "is_restricted      1535\n",
      "dtype: int64\n",
      "is_public        88.006855\n",
      "is_private       11.717410\n",
      "is_restricted     0.275735\n",
      "dtype: float64\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visibility of functions\n",
    "unsafe_function_definitions[\"is_private\"] = unsafe_function_definitions.visibility == \"Private\"\n",
    "unsafe_function_definitions[\"is_restricted\"] = unsafe_function_definitions.visibility.isin(set([\"Crate\", \"Restricted\"]))\n",
    "unsafe_function_definitions[\"is_public\"] = unsafe_function_definitions.visibility == \"Public\"\n",
    "\n",
    "unsafe_function_definitions[\"is_private_or_restricted\"] = unsafe_function_definitions.is_private | unsafe_function_definitions.is_restricted\n",
    "\n",
    "feature_names = [\n",
    "    \"is_private\",\n",
    "    \"is_restricted\",\n",
    "    \"is_public\",\n",
    "]\n",
    "\n",
    "num_trait_methods = sum(unsafe_function_definitions.is_trait_item)\n",
    "num_non_trait_methods = sum(~unsafe_function_definitions.is_trait_item)\n",
    "\n",
    "# Validation: there are no trait methods left\n",
    "assert sum([sum(unsafe_function_definitions[f]) for f in feature_names]) == num_non_trait_methods\n",
    "assert len(unsafe_function_definitions) == num_trait_methods + num_non_trait_methods\n",
    "\n",
    "print(\"Total number of unsafe functions:\", len(unsafe_function_definitions))\n",
    "print(\"Total number of trait-method unsafe functions: {} ({:%})\".format(\n",
    "    num_trait_methods, num_trait_methods/len(unsafe_function_definitions))\n",
    ")\n",
    "print(\"Total number of non-trait-method unsafe functions:{} ({:%})\".format(\n",
    "    num_non_trait_methods, num_non_trait_methods/len(unsafe_function_definitions))\n",
    ")\n",
    "\n",
    "features_frequency = unsafe_function_definitions[feature_names].apply(sum).sort_values(ascending=False)\n",
    "features_frequency.plot.bar(title=\"Visibility of unsafe functions\")\n",
    "print(features_frequency)\n",
    "print(features_frequency/num_non_trait_methods * 100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Follow-up 1\n",
    "\n",
    "> The results are depicted in Fig. 9. Based on this updated metric, we get a clearer picture: most crates (78.5%) have either all or none of their unsafe functions declared public. In particular, 34.7% of all crates seem to be well encapsulated: they declare unsafe functions but none of them are visible from the outside.\n",
    "\n",
    "Fig.9:\n",
    "\n",
    "![Fig.9](pics/rq4-distribution.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "public_unsafe_function_counts = dict(\n",
    "    unsafe_function_definitions.query('is_public').build.value_counts().iteritems())\n",
    "non_public_unsafe_function_counts = dict(unsafe_function_definitions.query('is_private_or_restricted').build.value_counts().iteritems())\n",
    "unsafe_function_proportions = []\n",
    "for (_, row) in selected_builds.iterrows():\n",
    "    build = row['build']\n",
    "    public_count = public_unsafe_function_counts.get(build) or 0\n",
    "    non_public_count = non_public_unsafe_function_counts.get(build) or 0\n",
    "    count = (non_public_count+public_count)\n",
    "    if count > 0:\n",
    "        row[\"public_count\"] = public_count\n",
    "        row[\"non_public_count\"] = non_public_count\n",
    "        row[\"total\"] = count\n",
    "        row[\"public_proportion\"] = public_count/count\n",
    "        unsafe_function_proportions.append(row)\n",
    "unsafe_function_proportions = pd.DataFrame(data=unsafe_function_proportions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = unsafe_function_proportions.public_proportion.plot.hist(\n",
    " bins=100,\n",
    " linewidth=4,\n",
    " fontsize=26,\n",
    ")\n",
    "ax.title.set_size(20)\n",
    "ax.set_xlabel(\"Proportion of unsafe functions declared public\", fontsize=32)\n",
    "ax.set_ylabel(\"Number of crates\", fontsize=32)\n",
    "ax.set_axisbelow(True)\n",
    "ax.grid(True, linestyle='dotted')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Moreover, 43.8% of crates declare all of their unsafe functions public;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1486 (34.7196261682243%) builds (of 4280) with unsafe functions have no public unsafe functions\n",
      "1873 (43.76168224299065%) builds (of 4280) with unsafe functions have only public unsafe function\n",
      "3359 (78.48130841121495%) builds (of 4280) with unsafe functions have either all unsafe function private or public\n"
     ]
    }
   ],
   "source": [
    "# Proportions:\n",
    "non_public_dominated_builds = set(\n",
    "    unsafe_function_proportions.query(\"public_count == 0\").build)\n",
    "non_public_dominated_builds_count = len(non_public_dominated_builds)\n",
    "print(\"{} ({}%) builds (of {}) with unsafe functions have no public unsafe functions\".format(\n",
    "    non_public_dominated_builds_count,\n",
    "    100*non_public_dominated_builds_count/len(unsafe_function_proportions),\n",
    "    len(unsafe_function_proportions)\n",
    "))\n",
    "public_dominated_builds = set(\n",
    "    unsafe_function_proportions.query(\"non_public_count == 0\").build)\n",
    "public_dominated_builds_count = len(public_dominated_builds)\n",
    "print(\"{} ({}%) builds (of {}) with unsafe functions have only public unsafe function\".format(\n",
    "    public_dominated_builds_count,\n",
    "    100*public_dominated_builds_count/len(unsafe_function_proportions),\n",
    "    len(unsafe_function_proportions)\n",
    "))\n",
    "assert len(unsafe_function_proportions.query(\"non_public_count == 0 and public_count == 0\")) == 0\n",
    "dominated_builds_count = non_public_dominated_builds_count + public_dominated_builds_count\n",
    "print(\"{} ({}%) builds (of {}) with unsafe functions have either all unsafe function private or public\".format(\n",
    "    dominated_builds_count,\n",
    "    100*dominated_builds_count/len(unsafe_function_proportions),\n",
    "    len(unsafe_function_proportions)\n",
    "))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Follow-up 2\n",
    "\n",
    "> more precisely, these crates contain 274434 (49.2%) unsafe functions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The public dominated tower contains 274434 (49.23642757185403%) from 557380 unsafe functions\n"
     ]
    }
   ],
   "source": [
    "# Unsafe functions from the public dominated tower.\n",
    "public_unsafe_functions = unsafe_function_definitions[\n",
    "    unsafe_function_definitions[\"build\"].isin(public_dominated_builds)\n",
    "].query(\"not is_trait_item\")\n",
    "assert len(public_unsafe_functions) > 0, len(public_unsafe_functions)\n",
    "assert len(public_unsafe_functions.query(\"not is_public\")) == 0, len(public_unsafe_functions.query(\"not is_public\"))\n",
    "public_unsafe_functions_count = len(public_unsafe_functions)\n",
    "print(\"The public dominated tower contains {} ({}%) from {} unsafe functions\".format(\n",
    "      public_unsafe_functions_count,\n",
    "      100*public_unsafe_functions_count/len(unsafe_function_definitions),\n",
    "      len(unsafe_function_definitions),\n",
    "))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> We found that 163,650 (59.6%) have foreign item ABI, which means that they are bindings of foreign items (most likely C functions). We also found that 571 (0.2%) functions have C ABI, which means that they can be called from C code and, therefore, it makes sense to have them public. The vast majority of the remaining functions (110,212 or 40.2%) have Rust ABI and, therefore, it is hard to automatically tell whether they are bindings or not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "def show_abi_counts(functions, total_count):\n",
    "    abi_counts = functions.abi.value_counts()\n",
    "    print(\"Counts:\")\n",
    "    print(abi_counts)\n",
    "    print()\n",
    "    print(\"Proportions:\")\n",
    "    print(abi_counts/len(functions) * 100)\n",
    "    print()\n",
    "    print(\"Of all:\")\n",
    "    print(abi_counts/total_count * 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Counts:\n",
      "ForeignItem    163650\n",
      "Rust           110212\n",
      "C                 571\n",
      "system              1\n",
      "Name: abi, dtype: int64\n",
      "\n",
      "Proportions:\n",
      "ForeignItem    59.631824\n",
      "Rust           40.159747\n",
      "C               0.208065\n",
      "system          0.000364\n",
      "Name: abi, dtype: float64\n",
      "\n",
      "Of all:\n",
      "ForeignItem    59.631824\n",
      "Rust           40.159747\n",
      "C               0.208065\n",
      "system          0.000364\n",
      "Name: abi, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# Show ABIs of the public dominated tower\n",
    "show_abi_counts(public_unsafe_functions, public_unsafe_functions_count)\n",
    "# Keep only Rust ABI\n",
    "public_rust_unsafe_functions = public_unsafe_functions.query(\"abi == 'Rust'\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Follow-up 3\n",
    "\n",
    "> Therefore, we checked the meta information of the crates that contain these functions and found that 9,642 (3.5%) are assigned to categories that indicate them as crates that wrap system libraries and 49,363 (18.0%) are assigned to categories related to embedded programming."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "unsafe_function_spans = utils.load_unsafe_function_spans()\n",
    "categories = utils.load_categories()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# We collected this list by manually checking all categories.\n",
    "ffi = [\n",
    "    \"external-ffi-bindings\",\n",
    "    \"ffi\",\n",
    "    \"libffi\",\n",
    "    \"ffi-bindings\",\n",
    "    \"cffi\",\n",
    "    \"ffistrings\",\n",
    "    \"api bindings\",\n",
    "    \"api-bindings\",\n",
    "    \"bindings\",\n",
    "    \"ffi-bindings\",\n",
    "    \"api-bindings\",\n",
    "]\n",
    "ffi_builds = set(categories[categories['category'].isin(ffi)].build)\n",
    "\n",
    "embedded = [\n",
    "    \"arm\",\n",
    "    \"cortex-m\",\n",
    "    \"embedded\",\n",
    "    \"ESP32\",\n",
    "    \"Embedded development\",\n",
    "]\n",
    "embedded_builds = set(categories[categories['category'].isin(embedded)].build)\n",
    "def check(name, suffix):\n",
    "    return name.lower().replace('-', '_').endswith(suffix)\n",
    "sys_package_names = set([\n",
    "    x for x in unsafe_function_spans.package_name.unique()\n",
    "    if check(x, \"_sys\") or\n",
    "       check(x, \"_bindgen\") or\n",
    "       check(x, \"_ffi\") or\n",
    "       check(x, \"_bindings\")\n",
    "])\n",
    "\n",
    "def report_counts(fns, total_count):\n",
    "    ffi_build_count = len(fns[\n",
    "        fns['build'].isin(ffi_builds)\n",
    "        | fns['package_name'].isin(sys_package_names)\n",
    "    ])\n",
    "    print('FFI:', ffi_build_count, ' or ', ffi_build_count/total_count*100, '%')\n",
    "\n",
    "    embedded_build_count = len(fns[fns['build'].isin(embedded_builds)])\n",
    "    print('embedded:', embedded_build_count,\n",
    "          ' or ', embedded_build_count/total_count*100, '%')\n",
    "\n",
    "    ffi_and_embedded_build_count = len(fns[\n",
    "        fns['build'].isin(embedded_builds) &\n",
    "        (fns['build'].isin(ffi_builds) |\n",
    "        fns['package_name'].isin(sys_package_names))\n",
    "    ])\n",
    "    print('embedded and FFI:', ffi_and_embedded_build_count,\n",
    "          ' or ', ffi_and_embedded_build_count/total_count*100, '%')\n",
    "\n",
    "    ffi_or_embedded_build_count = len(fns[\n",
    "        fns['build'].isin(embedded_builds) |\n",
    "        fns['build'].isin(ffi_builds) |\n",
    "        fns['package_name'].isin(sys_package_names)\n",
    "    ])\n",
    "    print('embedded or FFI:', ffi_or_embedded_build_count,\n",
    "          ' or ', ffi_or_embedded_build_count/total_count*100, '%')\n",
    "\n",
    "    fns = fns[\n",
    "        (~fns['build'].isin(ffi_builds)) &\n",
    "        (~fns['package_name'].isin(sys_package_names)) &\n",
    "        (~fns['build'].isin(embedded_builds))\n",
    "    ]\n",
    "    print(\"remaining:\", len(fns))\n",
    "    \n",
    "    return fns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FFI: 9642  or  3.5134130610638623 %\n",
      "embedded: 49363  or  17.987202751845615 %\n",
      "embedded and FFI: 4  or  0.0014575453478796359 %\n",
      "embedded or FFI: 59001  or  21.4991582675616 %\n",
      "remaining: 51211\n"
     ]
    }
   ],
   "source": [
    "public_rust_unsafe_function_spans = unsafe_function_spans[\n",
    "    unsafe_function_spans[\"build\"].isin(public_dominated_builds)\n",
    "].query(\"abi=='Rust' and visibility=='Public'\")\n",
    "\n",
    "assert len(public_rust_unsafe_functions) == len(public_rust_unsafe_function_spans)\n",
    "\n",
    "remaining_public_unsafe_functions = report_counts(\n",
    "    public_rust_unsafe_function_spans, public_unsafe_functions_count)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Follow-up 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Finally, we manually reviewed 30 crates from the remaining\n",
    "list that have most unsafe functions (in total 41,063 functions\n",
    "or 15.0%) and found that they either provide APIs to microcontrollers or OpenGL bindings. After our analysis we are left with\n",
    "only 10,148 functions (or 3.7%) that are public and which may not be from the crates that provide bindings."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Review comments:\n",
    "\n",
    "* `gdnative-*` – Rust bindings for GDNative.\n",
    "* Packages with unpronouncable names typically provide APIs for some microcontrolers: `cc13x2-cc26x2-pac`, `cc13x2-cc26x2-hal`, `oxcc-stm32f767`, `stm32f469xx`, `stm32wb-pac`, `cc2538`, `xmc4800`, `xmc4700`, `efr32xg1`, `xmc4500`, `alt-sam3x8e`, `nrf52840-pac`, `stm32l432-pac`, `stm32l4x2-pac`, `xmc4400`, `lpc177x_8x`, `xmc4300`, `stm32f103xx`, `xmc4100`, `xmc4200`, `nrf52832-pac`, `nrf52`, `stm32f072x-memory-map`, `atsame54p20a`.\n",
    "  * All `xmc*` packages are generated by the same author.\n",
    "* `gleam`, `sparkle`, `epoxy`, `grr`, `glow`, `gfx_gl`, `ogl33`, `gles30` – OpenGL bindings\n",
    "* `core`, `core_arch` – part of the standard library\n",
    "* `mallumo-gls` – mostly safe OpenGl bindings, but also exposes raw bindings in the `raw` module\n",
    "* `rg3d`, `surfman` – OpenGL bindings. Note: public unsafe functions are inside a private module.\n",
    "* `ioctls` – IO control bindings.\n",
    "* `ico_math` – exposes raw SSE extensions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 4267 https://crates.io/crates/cc13x2-cc26x2-pac\n",
      "1 4267 https://crates.io/crates/cc13x2-cc26x2-hal\n",
      "2 3149 https://crates.io/crates/oxcc-stm32f767\n",
      "3 2644 https://crates.io/crates/stm32f469xx\n",
      "4 1771 https://crates.io/crates/sparkle\n",
      "5 1737 https://crates.io/crates/stm32wb-pac\n",
      "6 1625 https://crates.io/crates/epoxy\n",
      "7 1325 https://crates.io/crates/cc2538\n",
      "8 1296 https://crates.io/crates/xmc4800\n",
      "9 1295 https://crates.io/crates/xmc4700\n",
      "10 1228 https://crates.io/crates/efr32xg1\n",
      "11 1168 https://crates.io/crates/xmc4500\n",
      "12 1146 https://crates.io/crates/alt-sam3x8e\n",
      "13 1072 https://crates.io/crates/nrf52840-pac\n",
      "14 1068 https://crates.io/crates/stm32l432-pac\n",
      "15 1068 https://crates.io/crates/stm32l4x2-pac\n",
      "16 1060 https://crates.io/crates/xmc4400\n",
      "17 1017 https://crates.io/crates/lpc177x_8x\n",
      "18 984 https://crates.io/crates/xmc4300\n",
      "19 881 https://crates.io/crates/stm32f103xx\n",
      "20 873 https://crates.io/crates/xmc4200\n",
      "21 873 https://crates.io/crates/xmc4100\n",
      "22 867 https://crates.io/crates/nrf52832-pac\n",
      "23 866 https://crates.io/crates/nrf52\n",
      "24 825 https://crates.io/crates/mallumo-gls\n",
      "25 703 https://crates.io/crates/gfx_gl\n",
      "26 589 https://crates.io/crates/ioctls\n",
      "27 525 https://crates.io/crates/atsame54p20a\n",
      "28 437 https://crates.io/crates/atsamd51j20a\n",
      "29 437 https://crates.io/crates/atsamd51j19a\n",
      "41063\n",
      "The number of functions: 41063\n",
      "Percentage (of all non-trait-method functions): 14.962796154995372 %\n",
      "Remaining: 10148\n",
      "Percentage (of all non-trait-method functions): 3.6977925475706366 %\n"
     ]
    }
   ],
   "source": [
    "s = 0\n",
    "array = []\n",
    "for n, v in remaining_public_unsafe_functions.package_name.value_counts().iteritems():\n",
    "    s += v\n",
    "    print(len(array), v, 'https://crates.io/crates/' + n)\n",
    "    array.append(n)\n",
    "    if len(array) == 30:\n",
    "        print(s)\n",
    "        break\n",
    "\n",
    "print(\"The number of functions:\", s)\n",
    "print(\"Percentage (of all non-trait-method functions):\", s/public_unsafe_functions_count*100, \"%\")\n",
    "print(\"Remaining:\", len(remaining_public_unsafe_functions)-s)\n",
    "print(\"Percentage (of all non-trait-method functions):\",\n",
    "      (len(remaining_public_unsafe_functions)-s)/public_unsafe_functions_count*100, \"%\")"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}