update_render.sh

#!/bin/bash
# -----------------------------------------------------------------------------
# update_render.sh
#
# Copyright (C) 2018-2025  Andy Townsend
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
# -----------------------------------------------------------------------------
# Designed to update rendering database and related styles to latest version.
# Note that it won't run on e.g. an NTFS file system, and makes a number of
# assumptions about where things are.
# -----------------------------------------------------------------------------
#
# The local user account we are using.
# "local_filesystem_user" is whichever non-root account is used to fetch from
# github.
# On Debian 11 or above and Ubuntu 21.04 and above,
# "local_renderd_user" will probably be "_renderd"
#
# Expected values for "local_database" are "gis" and "=gis3".
# If the data is loaded into gis it'll go live immediately after script completion.
# If the data is loaded into gis3 it can be tested and then taken live with make_gis3_live.sh
#
local_filesystem_user=ajtown
local_renderd_user=_renderd
local_database=gis3
#
# First things first - define some shared functions
#
reinstate_crontabs()
{
    crontab -u $local_renderd_user local_renderd_user_crontab_safe.$$
    mv osm_ldp1 /etc/cron.d/
    mv osm_ldp2 /etc/cron.d/
    mv osm_ldp3 /etc/cron.d/
}

final_tidy_up()
{
    rm last_modified1.$$ last_modified2.$$ last_modified3.$$ last_modified4.$$
    rm update_render.running
}

m_error_01()
{
    reinstate_crontabs
    final_tidy_up
    date | mail -s "Database reload FAILED on `hostname`, previous database ${local_database} intact" ${local_filesystem_user}
    exit 1
}

m_error_02()
{
    reinstate_crontabs
    final_tidy_up
    date | mail -s "Database reload FAILED on `hostname`, previous database ${local_database} lost" ${local_filesystem_user}
    exit 1
}

#
# Before anything else, is an update script for the previous database still running?
#
if [[ -f /var/cache/renderd/pyosmium.gis/call_pyosmium.running ]]
then
    echo "call_pyosmium still running; /var/cache/renderd/pyosmium.gis/call_pyosmium.running exists"
    exit 1
fi
#
# Next, is another copy of the script already running?
#
cd /home/${local_filesystem_user}/data
if test -e update_render.running
then
    echo update_render.running exists so exiting
    exit 1
else
    touch update_render.running
fi
# -----------------------------------------------------------------------------
# This script works with several different geographic regions.  
# Areas within Great Britain can be processed into CY, GD and EN speaking 
# areas and are handled by the "1" versions of e.v.s.
# Ireland, if needed, needs no language processing and is is handled by the 
# "2" versions of e.v.s.
# The results are then merged together.
#
# If 2 GB counties are merged, they must be from the same Geofabrik extract, 
# otherwise the "osmium merge" will fail, because some node in the GB 
# coastline will have one version in one file and another version in another.
# Any coastal English county (and Wales and Scotland) will include nodes in 
# the coastline which are part of the island of GB, which will be dragged in
# in full to both files.
# "Using current data" (see comment below) will work if something like the
# "Isle of Man" or "Ireland" is used as the second file.
# ----------------------------------------------------------------------------
# What's the first file that we are interested in?
#
#file_prefix1=europe
#file_page1=https://download.geofabrik.de/${file_prefix1}.html
#file_url1=https://download.geofabrik.de/${file_prefix1}-latest.osm.pbf
#
#file_prefix1=britain-and-ireland
file_prefix1=united-kingdom
#file_prefix1=ireland-and-northern-ireland
file_page1=https://download.geofabrik.de/europe/${file_prefix1}.html
file_url1=https://download.geofabrik.de/europe/${file_prefix1}-latest.osm.pbf
#
#file_prefix1=england
#file_prefix1=scotland
#file_prefix1=wales
#file_page1=https://download.geofabrik.de/europe/united-kingdom/${file_prefix1}.html
#file_url1=https://download.geofabrik.de/europe/united-kingdom/${file_prefix1}-latest.osm.pbf
#
#file_prefix1=bedfordshire
#file_prefix1=berkshire
#file_prefix1=bristol
#file_prefix1=buckinghamshire
#file_prefix1=cambridgeshire
#file_prefix1=cheshire
#file_prefix1=cornwall
#file_prefix1=cumbria
#file_prefix1=derbyshire
#file_prefix1=devon
#file_prefix1=dorset
#file_prefix1=durham
#file_prefix1=east-sussex
#file_prefix1=east-yorkshire-with-hull
#file_prefix1=essex
#file_prefix1=gloucestershire
#file_prefix1=greater-london
#file_prefix1=greater-manchester
#file_prefix1=hampshire
#file_prefix1=herefordshire
#file_prefix1=hertfordshire
#file_prefix1=isle-of-wight
#file_prefix1=kent
#file_prefix1=lancashire
#file_prefix1=leicestershire
#file_prefix1=lincolnshire
#file_prefix1=merseyside
#file_prefix1=norfolk
#file_prefix1=north-yorkshire
#file_prefix1=northamptonshire
#file_prefix1=northumberland
#file_prefix1=nottinghamshire
#file_prefix1=oxfordshire
#file_prefix1=rutland
#file_prefix1=shropshire
#file_prefix1=somerset
#file_prefix1=south-yorkshire
#file_prefix1=staffordshire
#file_prefix1=suffolk
#file_prefix1=surrey
#file_prefix1=tyne-and-wear
#file_prefix1=warwickshire
#file_prefix1=west-midlands
#file_prefix1=west-sussex
#file_prefix1=west-yorkshire
#file_prefix1=wiltshire
#file_prefix1=worcestershire
#file_page1=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix1}.html
#file_url1=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix1}-latest.osm.pbf
#
#file_prefix1=new-york
#file_prefix1=oregon
#file_page1=https://download.geofabrik.de/north-america/us/${file_prefix1}.html
#file_url1=https://download.geofabrik.de/north-america/us/${file_prefix1}-latest.osm.pbf
#
#file_prefix1=argentina
#file_page1=https://download.geofabrik.de/south-america/${file_prefix1}.html
#file_url1=https://download.geofabrik.de/south-america/${file_prefix1}-latest.osm.pbf
#
# What's the second file that we are interested in?
# Note that if this is commented out, also change the "merge" below to not use it.
#
file_prefix2=ireland-and-northern-ireland
#file_prefix2=isle-of-man
file_page2=https://download.geofabrik.de/europe/${file_prefix2}.html
file_url2=https://download.geofabrik.de/europe/${file_prefix2}-latest.osm.pbf
#file_prefix2=east-yorkshire-with-hull
#file_prefix2=rutland
#file_page2=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix2}.html
#file_url2=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix2}-latest.osm.pbf
#
# What's the third file that we are interested in?
# Note that if this is commented out, also change the "merge" below to not use it.
#
file_prefix3=isle-of-man
file_page3=https://download.geofabrik.de/europe/${file_prefix3}.html
file_url3=https://download.geofabrik.de/europe/${file_prefix3}-latest.osm.pbf
#file_prefix3=rutland
#file_prefix3=south-yorkshire
#file_page3=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix3}.html
#file_url3=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix3}-latest.osm.pbf
#
# What's the fourth file that we are interested in?
# Note that if this is commented out, also change the "merge" below to not use it.
#
file_prefix4=guernsey-jersey
file_page4=https://download.geofabrik.de/europe/${file_prefix4}.html
file_url4=https://download.geofabrik.de/europe/${file_prefix4}-latest.osm.pbf
#file_prefix4=rutland
#file_prefix4=west-yorkshire
#file_page4=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix4}.html
#file_url4=https://download.geofabrik.de/europe/united-kingdom/england/${file_prefix4}-latest.osm.pbf
#
# Remove some entries including the openstreetmap-tiles-update-expire one
# from the crontab.  Note that this matches a comment on the crontab line.
# The files are stored safely and restored at the end of the process.
#
crontab -u $local_renderd_user -l > local_renderd_user_crontab_safe.$$
grep -v "\#CONTROLLED BY update_render.sh" local_renderd_user_crontab_safe.$$ > local_renderd_user_crontab_new.$$
crontab -u $local_renderd_user local_renderd_user_crontab_new.$$
rm local_renderd_user_crontab_new.$$
#
# Also move some cron entries from "/etc/cron.d" out of the way
#
mv /etc/cron.d/osm_ldp1 .
mv /etc/cron.d/osm_ldp2 .
mv /etc/cron.d/osm_ldp3 .
#
# Next get the latest versions of each part of the map style.
#
# This is run from sudo without a connection to an authentication agent, 
# so it makes sense for the git config url to be "https" and the pushurl "git".  
# See https://stackoverflow.com/a/73836045/8145448
#
cd /home/${local_filesystem_user}/src/SomeoneElse-style
pwd
sudo -u ${local_filesystem_user} git pull
#
cd /home/${local_filesystem_user}/src/SomeoneElse-style-legend
pwd
sudo -u ${local_filesystem_user} git checkout generated_legend_pubs.osm
sudo -u ${local_filesystem_user} git pull
#
cd /home/${local_filesystem_user}/src/openstreetmap-carto-AJT
pwd
sudo -u ${local_filesystem_user} git pull
# -----------------------------------------------------------------------------
# We create 3 XML files:
#
# The normal mapnik.xml, which uses database gis.
# This is the main live tile layer and database.
#
# An identical mapnik3.xml, but which uses database gis3.
# Used for testing after database reload by setting "local_database" above to gis3, 
# and can then be made live later with "make_gis3_live.sh".
# There is normally no "gis2"; that is temporarily used by "make_gis3_live.sh".
#
# Also, a "novispaths" mapnik.xml, tile layer ajt5, which uses database gis, 
# which is based on the very simple layers at
# https://github.com/SomeoneElseOSM/openstreetmap-carto-AJT/tree/master/novispaths
# This is a separate tile layer so that it can be used as an overlay.
#
# For completeness, the tile layers / databases used are:
# gis   ajt   Main live database and tile layer
# gis2  n/a   Temporary only
# gis3  ajt3  For testing
# gis4  ajt4  https://github.com/SomeoneElseOSM/floodedmap
# gis   ajt5  No Vis Paths
# gis6  ajt6  https://github.com/SomeoneElseOSM/Boundary_Scripts
# -----------------------------------------------------------------------------
carto project.mml > mapnik.xml.new
sed "s/\[gis\]/[gis3]/" mapnik.xml.new > mapnik3.xml
#
if [ "local_database" = "gis" ]
then
    mv mapnik.xml mapnik.xml.old
    mv mapnik.xml.new mapnik.xml
fi
#
cd novispaths
carto project.mml > mapnik.xml
cd ..
#
# How much disk space are we currently using?
#
df
cd /home/${local_filesystem_user}/data
#
# When was the first target file last modified?
#
if [ "$1" = "current" ]
then
    echo "Using current data"
    ls -t | grep "${file_prefix1}_" | head -1 | sed "s/${file_prefix1}_//" | sed "s/.osm.pbf//" > last_modified1.$$
else
    wget $file_page1 -O file_page1.$$
    grep " and contains all OSM data up to " file_page1.$$ | sed "s/.*and contains all OSM data up to //" | sed "s/. File size.*//" > last_modified1.$$
    rm file_page1.$$
fi
#
file_extension1=`cat last_modified1.$$`
#
if test -e ${file_prefix1}_${file_extension1}.osm.pbf
then
    echo "File1 already downloaded"
else
    wget $file_url1 -O ${file_prefix1}_${file_extension1}.osm.pbf
fi
#
# When was the second target file last modified?
#
if [ "$1" = "current" ]
then
    ls -t | grep "${file_prefix2}_" | head -1 | sed "s/${file_prefix2}_//" | sed "s/.osm.pbf//" > last_modified2.$$
else
    wget $file_page2 -O file_page2.$$
    grep " and contains all OSM data up to " file_page2.$$ | sed "s/.*and contains all OSM data up to //" | sed "s/. File size.*//" > last_modified2.$$
    rm file_page2.$$
fi
#
file_extension2=`cat last_modified2.$$`
#
if test -e ${file_prefix2}_${file_extension2}.osm.pbf
then
    echo "File2 already downloaded"
else
    wget $file_url2 -O ${file_prefix2}_${file_extension2}.osm.pbf
fi
#
# When was the third target file last modified?
#
if [ "$1" = "current" ]
then
    ls -t | grep "${file_prefix3}_" | head -1 | sed "s/${file_prefix3}_//" | sed "s/.osm.pbf//" > last_modified3.$$
else
    wget $file_page3 -O file_page3.$$
    grep " and contains all OSM data up to " file_page3.$$ | sed "s/.*and contains all OSM data up to //" | sed "s/. File size.*//" > last_modified3.$$
    rm file_page3.$$
fi
#
file_extension3=`cat last_modified3.$$`
#
if test -e ${file_prefix3}_${file_extension3}.osm.pbf
then
    echo "File3 already downloaded"
else
    wget $file_url3 -O ${file_prefix3}_${file_extension3}.osm.pbf
fi
#
# When was the fourth target file last modified?
#
if [ "$1" = "current" ]
then
    ls -t | grep "${file_prefix4}_" | head -1 | sed "s/${file_prefix4}_//" | sed "s/.osm.pbf//" > last_modified4.$$
else
    wget $file_page4 -O file_page4.$$
    grep " and contains all OSM data up to " file_page4.$$ | sed "s/.*and contains all OSM data up to //" | sed "s/. File size.*//" > last_modified4.$$
    rm file_page4.$$
fi
#
file_extension4=`cat last_modified4.$$`
#
if test -e ${file_prefix4}_${file_extension4}.osm.pbf
then
    echo "File4 already downloaded"
else
    wget $file_url4 -O ${file_prefix4}_${file_extension4}.osm.pbf
fi
# -----------------------------------------------------------------------------
# Optionally stop rendering altogether (both renderd and apache2) 
# er to free up memory
# Alternatively, just restart renderd to reduce memory use.
# -----------------------------------------------------------------------------
/etc/init.d/renderd restart
echo ${file_prefix1} ${file_prefix2} ${file_prefix3} ${file_prefix4}
#/etc/init.d/apache2 stop
#
# In File1,
# Welsh, English and Scottish names need to be converted to "cy or en", "en" and "gd or en" respectively.
# First, convert a Welsh name portion into Welsh
#
if osmium extract --polygon /home/${local_filesystem_user}/src/SomeoneElse-style/welsh_areas.geojson ${file_prefix1}_${file_extension1}.osm.pbf -O -o welshlangpart_${file_extension1}_before.pbf
then
    echo Welsh Extract OK
else
    echo Welsh Extract Error
    m_error_01
fi

#
if /home/${local_filesystem_user}/src/osm-tags-transform/build/src/osm-tags-transform -c /home/${local_filesystem_user}/src/SomeoneElse-style/transform_cy.lua /home/${local_filesystem_user}/data/welshlangpart_${file_extension1}_before.pbf -O -o /home/${local_filesystem_user}/data/welshlangpart_${file_extension1}_after.pbf
then
    echo Welsh Transform OK
else
    echo Welsh Transform Error
    m_error_01
fi

#
# Likewise, Scots Gaelic
#
if osmium extract --polygon /home/${local_filesystem_user}/src/SomeoneElse-style/scotsgd_areas.geojson ${file_prefix1}_${file_extension1}.osm.pbf -O -o scotsgdlangpart_${file_extension1}_before.pbf
then
    echo Scots Gaelic Extract OK
else
    echo Scots Gaelic Extract Error
    m_error_01
fi

#
if /home/${local_filesystem_user}/src/osm-tags-transform/build/src/osm-tags-transform -c /home/${local_filesystem_user}/src/SomeoneElse-style/transform_gd.lua /home/${local_filesystem_user}/data/scotsgdlangpart_${file_extension1}_before.pbf -O -o /home/${local_filesystem_user}/data/scotsgdlangpart_${file_extension1}_after.pbf
then
    echo Scots Gaelic Transform OK
else
    echo Scots Gaelic Transform Error
    m_error_01
fi

#
# Unlike when using osmosis, which merges in a predictable way, 
# with osmium we have to explicitly extract the "English" part before conversion.
# The "English" geojson is a large multipolygon with the "Welsh" and "ScotsGD" areas as holes
# (using the exact same co-ordinates).
#
if osmium extract --polygon /home/${local_filesystem_user}/src/SomeoneElse-style/english_areas.geojson ${file_prefix1}_${file_extension1}.osm.pbf -O -o englangpart_${file_extension1}_before.pbf
then
    echo English Extract OK
else
    echo English Extract Error
    m_error_01
fi

#
if /home/${local_filesystem_user}/src/osm-tags-transform/build/src/osm-tags-transform -c /home/${local_filesystem_user}/src/SomeoneElse-style/transform_en.lua englangpart_${file_extension1}_before.pbf -O -o englangpart_${file_extension1}_after.pbf
then
    echo English Transform OK
else
    echo English Transform Error
    m_error_01
fi

#
# Note that "file2" through "file4" do not need splitting in this way; 
# "name" is used here.
# With "osmium merge" there is no way to merge so that cy and gd files 
# take precedence over the en one.
# See https://community.openstreetmap.org/t/osmium-merge-is-there-a-way-to-make-one-file-take-precedence/111017/2
#
# In most cases following the extracts above the data in each one is 
# mutually exclusive.  The exceptions are things (like the island of
# Great Britain) that are in both extracts.
#
if osmium merge ${file_prefix2}_${file_extension2}.osm.pbf ${file_prefix3}_${file_extension3}.osm.pbf ${file_prefix4}_${file_extension4}.osm.pbf englangpart_${file_extension1}_after.pbf welshlangpart_${file_extension1}_after.pbf scotsgdlangpart_${file_extension1}_after.pbf -O -o langs_${file_extension1}_merged.pbf
then
    echo Merge OK
else
    echo Merge Error
    m_error_01
fi

#
# Run osm2pgsql
#
mkdir -p /usr/local/share/lua/5.3/
cp /home/${local_filesystem_user}/src/SomeoneElse-style/shared_lua.lua /usr/local/share/lua/5.3/
#
if sudo -u ${local_renderd_user} osm2pgsql --create --slim -d ${local_database} -C 2500 --number-processes 2 -S /home/${local_filesystem_user}/src/openstreetmap-carto-AJT/openstreetmap-carto.style --multi-geometry --tag-transform-script /home/${local_filesystem_user}/src/SomeoneElse-style/style.lua langs_${file_extension1}_merged.pbf
then
    echo Database ${local_database} load OK
else
    echo Database ${local_database} load Error
    m_error_02
fi

#
if sudo -u ${local_renderd_user} osm2pgsql --append --slim -d ${local_database} -C 250 --number-processes 2 -S /home/${local_filesystem_user}/src/openstreetmap-carto-AJT/openstreetmap-carto.style --multi-geometry --tag-transform-script /home/${local_filesystem_user}/src/SomeoneElse-style/style.lua /home/${local_filesystem_user}/src/SomeoneElse-style-legend/legend_roads.osm
then
    echo Legend roads append OK
else
    echo Legend roads append Error
    m_error_02
fi

#
rm -f /home/${local_filesystem_user}/src/SomeoneElse-style-legend/generated_legend_pubs.osm

if sudo -u ${local_filesystem_user} osmium sort -o /home/${local_filesystem_user}/src/SomeoneElse-style-legend/generated_legend_pubs.osm /home/${local_filesystem_user}/src/SomeoneElse-style-legend/generated_legend_pub.osm
then
    echo Legend pubs sort OK
else
    echo Legend pubs sort Error
    m_error_02
fi

if sudo -u ${local_renderd_user} osm2pgsql --append --slim -d ${local_database} -C 250 --number-processes 2 -S /home/${local_filesystem_user}/src/openstreetmap-carto-AJT/openstreetmap-carto.style --multi-geometry --tag-transform-script /home/${local_filesystem_user}/src/SomeoneElse-style/style.lua /home/${local_filesystem_user}/src/SomeoneElse-style-legend/generated_legend_pubs.osm
then
    echo Legend pubs append OK
else
    echo Legend pubs append Error
    m_error_02
fi

#
date | mail -s "Database ${local_database} reload complete on `hostname`" ${local_filesystem_user}
#
# Tidy temporary files
#
rm welshlangpart_${file_extension1}_before.pbf welshlangpart_${file_extension1}_after.pbf englangpart_${file_extension1}_before.pbf englangpart_${file_extension1}_after.pbf scotsgdlangpart_${file_extension1}_before.pbf scotsgdlangpart_${file_extension1}_after.pbf langs_${file_extension1}_merged.pbf
#
# Note one of these next two sections only will be needed, depending on whether osmosis or osm2pgsql is used for updates
#
# Reinitialise updating (osmosis)
#
#rm -rf /var/lib/mod_tile/.osmosis.old
#mv /var/lib/mod_tile/.osmosis /var/lib/mod_tile/.osmosis.old
#sudo -u ${local_renderd_user} /home/${local_filesystem_user}/src/mod_tile/openstreetmap-tiles-update-expire ${file_extension1}
#
# Reinitialise updating (pyosmium)
#
rm -rf /var/cache/renderd/pyosmium.${local_database}.old
mv /var/cache/renderd/pyosmium.${local_database} /var/cache/renderd/pyosmium.${local_database}.old
mkdir /var/cache/renderd/pyosmium.${local_database}
chown ${local_renderd_user} /var/cache/renderd/pyosmium.${local_database}
sudo -u ${local_renderd_user} pyosmium-get-changes -D ${file_extension1} -f /var/cache/renderd/pyosmium.${local_database}/sequence.state -v
#
pandoc -s -f markdown -t html -H /home/${local_filesystem_user}/src/SomeoneElse-map/raster_header.html --metadata title="Changelog" /home/${local_filesystem_user}/src/SomeoneElse-style/changelog.md > /var/www/html/maps/map/changelog.html
#
pandoc -s -f markdown -t html -H /home/${local_filesystem_user}/src/SomeoneElse-map/raster_header.html --metadata title="About this map" /home/${local_filesystem_user}/src/SomeoneElse-map/about.md > /var/www/html/maps/map/about.html
#
/etc/init.d/renderd restart
/etc/init.d/apache2 restart
#
# Reinstate the crontabs
#
reinstate_crontabs
# 
# And final tidying up
#
final_tidy_up
#