analyze-stan-headers.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "76685b5a-490c-4a79-886f-30c4500a8de9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pathlib\n",
    "import re\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e7783d49-5440-4622-a535-543b69a2120b",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Generate prep.cpp file\n",
    "# g++ -E -std=c++1y -pthread -D_REENTRANT -Wno-sign-compare -Wno-ignored-attributes      -I ./stan/lib/stan_math/lib/tbb_2020.3/include    -O3 -I ./stan/src -I ./stan/lib/rapidjson_1.1.0/ -I ./stan/lib/stan_math/ -I ./stan/lib/stan_math/lib/eigen_3.4.0 -I ./stan/lib/stan_math/lib/boost_1.78.0 -I ./stan/lib/stan_math/lib/sundials_6.1.1/include -I ./stan/lib/stan_math/lib/sundials_6.1.1/src/sundials -fPIC    -DBOOST_DISABLE_ASSERTS          -c -x c++ -o prep.cpp test_models/logistic/logistic.hpp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "4d8d0a8d-5c6a-46e7-a2c8-0a7b171f1504",
   "metadata": {},
   "outputs": [],
   "source": [
    "content = pathlib.Path(\"/home/adr/.bridgestan/bridgestan-2.1.1/prep.cpp\").read_text()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "eff870cb-a362-4588-ada7-a60748e3ea69",
   "metadata": {},
   "outputs": [],
   "source": [
    "content = list(content.split(\"\\n\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "73930510-e174-47b3-9676-044616266593",
   "metadata": {},
   "outputs": [],
   "source": [
    "regex = re.compile('\\# ([0-9]*) \\\"([\\<\\>/a-zA-Z0-9\\._\\-]*)\\\"[ 0-9]*')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6b2582f1-77a7-472c-bf91-3e236b602cf4",
   "metadata": {},
   "outputs": [],
   "source": [
    "line_counts = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8aab3d93-389c-419e-948d-41b048d9e361",
   "metadata": {},
   "outputs": [],
   "source": [
    "current = None\n",
    "for line in content:\n",
    "    match = regex.match(line)\n",
    "    if not match and current:\n",
    "        line_counts[current] += 1\n",
    "        continue\n",
    "    if not match and not current:\n",
    "        assert False\n",
    "    _, current = match.groups()\n",
    "    line_counts.setdefault(current, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "eb261fe0-6f1b-4a2b-b055-27d3297cce70",
   "metadata": {},
   "outputs": [],
   "source": [
    "line_counts = pd.Series(line_counts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "699fbe1b-2cf0-48e5-a894-dc06c2e26766",
   "metadata": {},
   "outputs": [],
   "source": [
    "counts = line_counts.rename(\"count\").to_frame().rename_axis(index=\"file\").reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d0861a7e-c276-4e03-94db-aadb69a00ae6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3322"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of headers (more or less? What are those with zero lines?)\n",
    "len(line_counts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "09feba5d-a08f-4247-ae7b-a203b90b0bc8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2911"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(list(name for name, count in line_counts.items() if count > 0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "8a06fc04-c3d8-422f-8a9b-c18464c2c472",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "path\n",
       "/usr/include/bits/endian.h                             1\n",
       "/usr/include/bits/endianness.h                         1\n",
       "/usr/include/bits/environments.h                       1\n",
       "/usr/include/bits/cpu-set.h                            1\n",
       "/usr/include/bits/floatn-common.h                      1\n",
       "/usr/include/bits/floatn.h                             1\n",
       "/usr/include/bits/flt-eval-method.h                    1\n",
       "/usr/include/bits/fp-fast.h                            1\n",
       "/usr/include/bits/fp-logb.h                            1\n",
       "/usr/include/bits/getopt_core.h                        1\n",
       "/usr/include/bits/getopt_posix.h                       1\n",
       "/usr/include/bits/errno.h                              1\n",
       "/usr/lib/gcc/x86_64-pc-linux-gnu/13.1.1/include       10\n",
       "./stan/lib/stan_math/lib/sundials_6.1.1/include       21\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/unsupported      21\n",
       "./stan/lib/stan_math/lib/tbb_2020.3/include           28\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen           261\n",
       "./stan/lib/stan_math/stan/math/rev                   315\n",
       "./stan/lib/stan_math/stan/math/prim                  950\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost         1555\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of included files by source directory\n",
    "counts[\"path\"] = counts.file.str.split(\"/\").str.slice(0, 7).str.join(\"/\")\n",
    "counts.groupby(\"path\")[\"count\"].count().sort_values().tail(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "68caec11-b60a-4012-ac5c-77cfd5847276",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "path\n",
       "/usr/include/features.h                       621\n",
       "/usr/include/bits/confname.h                  658\n",
       "/usr/include/unistd.h                         660\n",
       "./stan/src/stan/io/deserializer.hpp           711\n",
       "/usr/include/pthread.h                       1068\n",
       "/usr/include/wchar.h                         1164\n",
       "./stan/src/stan/model/model_header.hpp       1865\n",
       "/usr/include/stdlib.h                        2298\n",
       "./stan/src/stan/model/indexing               2525\n",
       "./stan/lib/stan_math/lib/sundials_6.1.1      2608\n",
       "/usr/include/bits/mathcalls.h                2670\n",
       "/usr/include/libintl.h                       3017\n",
       "/usr/include/errno.h                         4636\n",
       "./stan/lib/stan_math/lib/tbb_2020.3          5363\n",
       "/usr/lib/gcc/x86_64-pc-linux-gnu/13.1.1      8326\n",
       "/usr/include/wctype.h                        8507\n",
       "/usr/include/sys/single_threaded.h          10620\n",
       "./stan/lib/stan_math/stan/math              83223\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0       103128\n",
       "./stan/lib/stan_math/lib/boost_1.78.0      189731\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of lines included from different directories\n",
    "counts[\"path\"] = counts.file.str.split(\"/\").str.slice(0, 6).str.join(\"/\")\n",
    "counts.groupby(\"path\")[\"count\"].sum().sort_values().tail(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "21a1184c-2cc7-423d-b2b2-f9c2c1a2f9eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "path\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/fusion/sequence                    2059\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/QR                              2062\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/LU                              2070\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/OrderingMethods                 2077\n",
       "./stan/lib/stan_math/stan/math/prim/core/init_threadpool_tbb.hpp               2150\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/math/policies                      2189\n",
       "/usr/include/stdlib.h                                                          2298\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/function_types/detail              2324\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/SVD                             2462\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/type_traits/detail                 2556\n",
       "/usr/include/bits/mathcalls.h                                                  2670\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/SparseLU                        2735\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/unsupported/Eigen/src                     2774\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/units/detail                       2965\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/fusion/functional                  2972\n",
       "/usr/include/libintl.h                                                         3017\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/lexical_cast/detail                3088\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/numeric/conversion                 3305\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/fusion/algorithm                   3694\n",
       "./stan/lib/stan_math/stan/math/prim/functor/hcubature.hpp                      3727\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/Eigenvalues                     3931\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/IterativeLinearSolvers          3982\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/fusion/view                        4278\n",
       "./stan/lib/stan_math/stan/math/prim/meta/holder.hpp                            4379\n",
       "/usr/lib/gcc/x86_64-pc-linux-gnu/13.1.1/include/stddef.h                       4436\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/Geometry                        4529\n",
       "/usr/include/errno.h                                                           4636\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/math/tools                         5101\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/mpl/aux_                           6230\n",
       "./stan/lib/stan_math/stan/math/rev/core/init_chainablestack.hpp                6331\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/SparseCore                          8127\n",
       "/usr/include/wctype.h                                                          8507\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/SparseCore                     10142\n",
       "/usr/include/sys/single_threaded.h                                            10620\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/Core                               11181\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/function/function_template.hpp    11379\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/numeric/odeint                    14368\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/math/special_functions            31605\n",
       "./stan/lib/stan_math/lib/boost_1.78.0/boost/numeric/ublas                     35105\n",
       "./stan/lib/stan_math/lib/eigen_3.4.0/Eigen/src/Core                           42699\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Same, just a bit more detailed\n",
    "counts[\"path\"] = counts.file.str.split(\"/\").str.slice(0, 9).str.join(\"/\")\n",
    "counts.groupby(\"path\")[\"count\"].sum().sort_values().tail(40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b70074c3-3e59-4912-bc3a-29ff7786d285",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dev",
   "language": "python",
   "name": "dev"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}