Rework download area support (#908)

Closes #904
* Make all data-related targets like `download*`, `import-osm`, `import-borders`, and `generate-tiles` into `area`-aware -- making it possible for multiple data files to coexist inside the `./data` dir.
* Add `make download area=... [url=...]` command to automatically download any kind of area by checking Geofabrik, BBBike, and OSM.fr, optionally from a custom URL. Supports `area=planet` too.
* Do not re-download area with `make download-*` if it already exists.
* Automatically rename `<area>-latest.osm.pbf` into `<area>.osm.pbf`
* If `area=...` parameter is not given to `make`, see if there is exactly one `*.osm.pbf` file, and if so, use `*` as the `area`.
* Configure many variables in the .env file, overriding the defaults in tools
* If `<area>.osm.pbf` exists, but `<area>.dc-config.pbf` is missing, generate it using `download-osm make-dc` command.

Also:
* closes #614
* closes #647
* partially addresses #261
pull/916/head
Yuri Astrakhan 2020-06-03 15:37:45 -04:00 zatwierdzone przez GitHub
rodzic fca53ef0ee
commit 1486b7e0cb
Nie znaleziono w bazie danych klucza dla tego podpisu
ID klucza GPG: 4AEE18F83AFDEB23
6 zmienionych plików z 184 dodań i 85 usunięć

11
.env
Wyświetl plik

@ -24,3 +24,14 @@ FILTER_MAPNIK_OUTPUT=1
# Some area data like openstreetmap.fr can contain invalid references
# that must be cleaned up before using it for borders -- set it to true.
BORDERS_CLEANUP=false
# The current setup assumes this file is placed inside the data/ dir
MBTILES_FILE=tiles.mbtiles
# This is the current repl_config.json location, pre-configured in the tools Dockerfile
# Makefile and quickstart replace it with the dynamically generated one, but we keep it here in case some other method is used to run.
IMPOSM_CONFIG_FILE=/usr/src/app/config/repl_config.json
# import-borders temp files - set them here to defaults, and override in the Makefile based on the area
BORDERS_CLEANUP_FILE=data/borders/cleanup.pbf
BORDERS_PBF_FILE=data/borders/filtered.pbf
BORDERS_CSV_FILE=data/borders/lines.csv

Wyświetl plik

@ -173,7 +173,7 @@ jobs:
create_db() {
make clean
make init-dirs
cp ../ci_cache/perf-test-areas-latest.osm.pbf data/perf-test-areas-latest.osm.pbf
cp ../ci_cache/perf-test-areas-latest.osm.pbf data/perf-test-areas.osm.pbf
make destroy-db
make all
make start-db

163
Makefile
Wyświetl plik

@ -13,9 +13,6 @@ DC_OPTS ?= --rm -u $(shell id -u):$(shell id -g)
# If set to a non-empty value, will use postgis-preloaded instead of postgis docker image
USE_PRELOADED_IMAGE ?=
# If set, this file will be imported in the import-osm target
PBF_FILE?=
# Local port to use with postserve
PPORT ?= 8090
export PPORT
@ -55,6 +52,87 @@ endif
OMT_HOST := http://$(firstword $(subst :, ,$(subst tcp://,,$(DOCKER_HOST))) localhost)
#
# Determine area to work on
# If $(area) parameter is not set and data/*.osm.pbf finds only one file, use it as $(area).
# Otherwise all make targets requiring area param will show an error.
# Note: If there are no data files, and user calls make download area=... once,
# they will not need to use area= parameter after that because there will be just a single file.
#
# historically we have been using $(area) rather than $(AREA), so make both work
area ?= $(AREA)
# Ensure the $(AREA) param is set, or try to automatically determine it based on available data files
ifeq ($(strip $(area)),)
# if $area is not set. set it to the name of the *.osm.pbf file, but only if there is only one
data_files := $(wildcard data/*.osm.pbf)
ifneq ($(word 2,$(data_files)),)
AREA_ERROR := The 'area' parameter (or env var) has not been set, and there are more than one data/*.osm.pbf files. Set area to one of these IDs, or a new one: $(patsubst data/%.osm.pbf,'%',$(data_files))
else
ifeq ($(word 1,$(data_files)),)
AREA_ERROR := The 'area' parameter (or env var) has not been set, and there are no data/*.osm.pbf files
else
# Keep just the name of the data file, without the .osm.pbf extension
area := $(strip $(basename $(basename $(notdir $(data_files)))))
# Rename area-latest.osm.pbf to area.osm.pbf
# TODO: This if statement could be removed in a few months once everyone is using the file without the `-latest`?
ifneq ($(area),$(area:-latest=))
$(shell mv "data/$(area).osm.pbf" "data/$(area:-latest=).osm.pbf")
area := $(area:-latest=)
$(warning ATTENTION: File data/$(area)-latest.osm.pbf was renamed to $(area).osm.pbf.)
AREA_INFO := Detected area=$(area) based on the found data/$(area)-latest.osm.pbf (renamed to $(area).osm.pbf). Use 'area' parameter (or env var) to override.
else
AREA_INFO := Detected area=$(area) based on the found data/ pbf file. Use 'area' parameter (or env var) to override.
endif
endif
endif
endif
# If set, this file will be downloaded in download-osm and imported in the import-osm targets
PBF_FILE ?= data/$(area).osm.pbf
# For download-osm, allow URL parameter to download file from a given URL. Area param must still be provided.
ifneq ($(strip $(url)),)
DOWNLOAD_AREA := $(url)
else
DOWNLOAD_AREA := $(area)
endif
# import-borders uses these temp files during border parsing/import
export BORDERS_CLEANUP_FILE ?= data/borders/$(area).cleanup.pbf
export BORDERS_PBF_FILE ?= data/borders/$(area).filtered.pbf
export BORDERS_CSV_FILE ?= data/borders/$(area).lines.csv
# The file is placed into the $EXPORT_DIR=/export (mapped to ./data)
export MBTILES_FILE ?= $(area).mbtiles
MBTILES_LOCAL_FILE = data/$(MBTILES_FILE)
# Location of the dynamically-generated imposm config file
export IMPOSM_CONFIG_FILE ?= data/$(area).repl.json
# download-osm generates this file with metadata about the file
AREA_DC_CONFIG_FILE ?= data/$(area).dc-config.yml
ifeq ($(strip $(area)),)
define assert_area_is_given
@echo "ERROR: $(AREA_ERROR)"
@echo ""
@echo " make $@ area=<area-id>"
@echo ""
@echo "To download an area, use make download <area-id>"
@echo "To list downloadable areas, use make list-geofabrik and/or make list-bbbike"
@exit 1
endef
else
ifneq ($(strip $(AREA_INFO)),)
define assert_area_is_given
@echo "$(AREA_INFO)"
endef
endif
endif
#
# TARGETS
#
@ -78,6 +156,8 @@ help:
@echo "Hints for developers:"
@echo " make # build source code"
@echo " make list-geofabrik # list actual geofabrik OSM extracts for download"
@echo " make list-bbbike # list actual BBBike OSM extracts for download"
@echo " make download area=albania # download OSM data from any source and create config file"
@echo " make download-geofabrik area=albania # download OSM data from geofabrik.de and create config file"
@echo " make download-osmfr area=asia/qatar # download OSM data from openstreetmap.fr and create config file"
@echo " make download-bbbike area=Amsterdam # download OSM data from bbbike.org and create config file"
@ -105,7 +185,7 @@ help:
.PHONY: init-dirs
init-dirs:
@mkdir -p build/sql
@mkdir -p data
@mkdir -p data/borders
@mkdir -p cache
build/openmaptiles.tm2source/data.yml: init-dirs
@ -165,29 +245,51 @@ list-geofabrik: init-dirs
list-bbbike: init-dirs
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools download-osm list bbbike
#
# download, download-geofabrik, download-osmfr, and download-bbbike are handled here
# The --imposm-cfg will fail for some of the sources, but we ignore that error -- only needed for diff mode
#
OSM_SERVERS := geofabrik osmfr bbbike
ALL_DOWNLOADS := $(addprefix download-,$(OSM_SERVERS))
OSM_SERVER=$(patsubst download-%,%,$@)
ALL_DOWNLOADS := $(addprefix download-,$(OSM_SERVERS)) download
OSM_SERVER=$(patsubst download,,$(patsubst download-%,%,$@))
.PHONY: $(ALL_DOWNLOADS)
$(ALL_DOWNLOADS): init-dirs
ifeq ($(strip $(area)),)
@echo ""
@echo "ERROR: Unable to download an area if area is not given."
@echo "Usage:"
@echo " make download-$(OSM_SERVER) area=<area-id>"
@echo ""
$(if $(filter %-geofabrik,$@),@echo "Use make list-geofabrik to get a list of all available areas";echo "")
@exit 1
else
@echo "=============== download-$(OSM_SERVER) ======================="
@echo "Download area: $(area)"
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools bash -c \
'download-osm $(OSM_SERVER) $(area) \
@$(assert_area_is_given)
ifeq (,$(wildcard $(PBF_FILE)))
ifneq ($(strip $(url)),)
$(if $(OSM_SERVER),$(error url parameter can only be used with the 'make download area=... url=...'))
endif
@echo "Downloading $(area) into $(PBF_FILE) from $(if $(OSM_SERVER),$(OSM_SERVER),any source)"
@$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools bash -c ' \
download-osm $(OSM_SERVER) $(DOWNLOAD_AREA) \
--minzoom $$QUICKSTART_MIN_ZOOM \
--maxzoom $$QUICKSTART_MAX_ZOOM \
--make-dc /import/docker-compose-config.yml -- -d /import'
ls -la ./data/$(notdir $(area))*
--make-dc $(AREA_DC_CONFIG_FILE) \
--imposm-cfg $(IMPOSM_CONFIG_FILE) \
--output $(PBF_FILE) \
2>&1 \
| tee /tmp/download.out ; \
exit_code=$${PIPESTATUS[0]} ; \
if [[ "$$exit_code" != "0" ]]; then \
if grep -q "Imposm config file cannot be generated from this source" /tmp/download.out; then \
echo "WARNING: $(IMPOSM_CONFIG_FILE) could not be generated, but it is only needed to apply updates." ; \
else \
exit $$exit_code ; \
fi ; \
fi'
@echo ""
else
ifeq (,$(wildcard $(AREA_DC_CONFIG_FILE)))
@echo "Data file $(PBF_FILE) already exists, but the $(AREA_DC_CONFIG_FILE) is not, generating..."
@$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools bash -c ' \
download-osm make-dc $(PBF_FILE) \
--minzoom $$QUICKSTART_MIN_ZOOM \
--maxzoom $$QUICKSTART_MAX_ZOOM \
--make-dc $(AREA_DC_CONFIG_FILE) \
--id "$(area)"'
else
@echo "Data files $(PBF_FILE) and $(AREA_DC_CONFIG_FILE) already exists, skipping the download."
endif
endif
.PHONY: psql
@ -196,6 +298,7 @@ psql: start-db-nowait
.PHONY: import-osm
import-osm: all start-db-nowait
@$(assert_area_is_given)
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools sh -c 'pgwait && import-osm $(PBF_FILE)'
.PHONY: update-osm
@ -212,24 +315,28 @@ import-data: start-db
.PHONY: import-borders
import-borders: start-db-nowait
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools sh -c 'pgwait && import-borders $(PBF_FILE)'
@$(assert_area_is_given)
# If CSV borders file already exists, use it without re-parsing
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools sh -c \
'pgwait && import-borders $$([ -f "$(BORDERS_CSV_FILE)" ] && echo 'load' || echo 'import') $(PBF_FILE)'
.PHONY: import-sql
import-sql: all start-db-nowait
$(DOCKER_COMPOSE) run $(DC_OPTS) openmaptiles-tools sh -c 'pgwait && import-sql' | \
awk -v s=": WARNING:" '$$0~s{print; print "\n*** WARNING detected, aborting"; exit(1)} 1'
.PHONY: generate-tiles
ifneq ($(wildcard data/docker-compose-config.yml),)
DC_CONFIG_TILES:=-f docker-compose.yml -f ./data/docker-compose-config.yml
ifneq ($(wildcard $(AREA_DC_CONFIG_FILE)),)
DC_CONFIG_TILES := -f docker-compose.yml -f $(AREA_DC_CONFIG_FILE)
endif
.PHONY: generate-tiles
generate-tiles: all start-db
rm -rf data/tiles.mbtiles
echo "Generating tiles ..."; \
@$(assert_area_is_given)
@echo "Generating tiles into $(MBTILES_LOCAL_FILE) (will delete if already exists)..."
@rm -rf "$(MBTILES_LOCAL_FILE)"
$(DOCKER_COMPOSE) $(DC_CONFIG_TILES) run $(DC_OPTS) generate-vectortiles
@echo "Updating generated tile metadata ..."
$(DOCKER_COMPOSE) $(DC_CONFIG_TILES) run $(DC_OPTS) openmaptiles-tools \
mbtiles-tools meta-generate ./data/tiles.mbtiles ./openmaptiles.yaml --auto-minmax --show-ranges
mbtiles-tools meta-generate "$(MBTILES_LOCAL_FILE)" ./openmaptiles.yaml --auto-minmax --show-ranges
.PHONY: start-tileserver
start-tileserver: init-dirs

Wyświetl plik

@ -83,10 +83,10 @@ make
```
You can execute the following manual steps (for better understanding)
or use the provided `quickstart.sh` script.
or use the provided `quickstart.sh` script to automatically download and import given area. If area is not given, albania will be imported.
```
./quickstart.sh
./quickstart.sh <area>
```
### Prepare the Database
@ -103,10 +103,10 @@ Import external data from [OpenStreetMapData](http://osmdata.openstreetmap.de/),
make import-data
```
[Download OpenStreetMap data extracts](http://download.geofabrik.de/) and store the PBF file in the `./data` directory.
Download OpenStreetMap data extracts from any source like [Geofabrik](http://download.geofabrik.de/), and store the PBF file in the `./data` directory. Use `download-geofabrik`, `download-bbbike`, or `download-osmfr` for a specific source. Use `download area=planet` for the entire OSM dataset (very large). Note that if you have more than one `data/*.osm.pbf` file, every `make` command will require `area=...` parameter (or you can just `export area=...` first)
```bash
make download-geofabrik area=albania
make download area=albania
```
[Import OpenStreetMap data](https://github.com/openmaptiles/openmaptiles-tools/tree/master/docker/import-osm) with the mapping rules from

Wyświetl plik

@ -36,6 +36,10 @@ services:
MAKE_DC_VERSION: "2.3"
# Allow DIFF_MODE to be overwritten from shell
DIFF_MODE: ${DIFF_MODE}
# Which files to use during import-borders processing
BORDERS_CLEANUP_FILE: ${BORDERS_CLEANUP_FILE}
BORDERS_PBF_FILE: ${BORDERS_PBF_FILE}
BORDERS_CSV_FILE: ${BORDERS_CSV_FILE}
networks:
- postgres_conn
volumes:
@ -54,6 +58,10 @@ services:
networks:
- postgres_conn
env_file: .env
environment:
FILTER_MAPNIK_OUTPUT: ${FILTER_MAPNIK_OUTPUT}
MBTILES_NAME: ${MBTILES_FILE}
IMPOSM_CONFIG_FILE: ${IMPOSM_CONFIG_FILE}
generate-vectortiles:
image: "openmaptiles/generate-vectortiles:${TOOLS_VERSION}"
@ -64,10 +72,12 @@ services:
- postgres_conn
env_file: .env
environment:
FILTER_MAPNIK_OUTPUT: ${FILTER_MAPNIK_OUTPUT}
MBTILES_NAME: ${MBTILES_FILE}
IMPOSM_CONFIG_FILE: ${IMPOSM_CONFIG_FILE}
BBOX: ${BBOX}
MIN_ZOOM: ${MIN_ZOOM}
MAX_ZOOM: ${MAX_ZOOM}
FILTER_MAPNIK_OUTPUT: ${FILTER_MAPNIK_OUTPUT}
postserve:
image: "openmaptiles/openmaptiles-tools:${TOOLS_VERSION}"

Wyświetl plik

@ -46,26 +46,17 @@ else
fi
if [ $# -eq 0 ]; then
osm_area=albania # default test country
echo "No parameter - set area=$osm_area "
# default test area
export area=albania
echo "No parameter - set area=$area "
else
osm_area=$1
export area=$1
fi
if [ $# -eq 2 ]; then
osm_server=$2
else
if [[ ${osm_area} == [[:upper:]]* ]]; then
# Only bbbike area names are capitalized
osm_server=bbbike
else
# default OSM server
osm_server=geofabrik
fi
osm_server=$2
fi
pbf_file="./data/${osm_area##*/}-latest.osm.pbf"
## Min versions ...
MIN_COMPOSE_VER=1.7.1
MIN_DOCKER_VER=1.12.3
@ -117,8 +108,8 @@ echo " Start processing
echo "-------------------------------------------------------------------------------------"
echo "====> : OpenMapTiles quickstart! [ https://github.com/openmaptiles/openmaptiles ] "
echo " : This will be logged to the $log_file file (for debugging) and to the screen"
echo " : Area : $osm_area "
echo " : Download Server : $osm_server "
echo " : Area : $area "
echo " : Download Server : ${osm_server:-unset (automatic)} "
echo " : Preloaded Image : $USE_PRELOADED_IMAGE "
echo " : Git version : $(git rev-parse HEAD) "
echo " : Started : $STARTDATE "
@ -140,7 +131,7 @@ if [[ "$OSTYPE" == "linux-gnu" ]]; then
exit 1
fi
echo " : --- Memory, CPU info ---- "
mem=$( grep MemTotal /proc/meminfo | awk '{print $2}' | xargs -I {} echo "scale=4; {}/1024^2" | bc )
mem=$( grep MemTotal /proc/meminfo | awk '{print $2}' | xargs -I {} echo "scale=4; {}/1024^2" | bc )
echo "system memory (GB): ${mem}"
grep SwapTotal /proc/meminfo
echo "cpu number: $(grep -c processor /proc/cpuinfo) x $(grep "bogomips" /proc/cpuinfo | head -1)"
@ -168,27 +159,13 @@ make init-dirs
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Removing old MBTILES if exists ( ./data/*.mbtiles ) "
rm -f ./data/*.mbtiles
echo "====> : Removing old MBTILES if exists ( ./data/${area}.mbtiles ) "
rm -f "./data/${area}.mbtiles"
if [[ ! -f "${pbf_file}" || ! -f "./data/docker-compose-config.yml" ]]; then
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Downloading ${osm_area} from ${osm_server}..."
rm -rf ./data/*
make download-${osm_server} "area=${osm_area}"
else
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : The pbf file $pbf_file exists, we don't need to download!"
fi
if [ ! -f "${pbf_file}" ]; then
echo " "
echo "Missing $pbf_file , Download or Parameter error? "
exit 1
fi
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Downloading ${area} from ${osm_server:-any source}..."
make "download${osm_server:+-${osm_server}}"
echo " "
echo "-------------------------------------------------------------------------------------"
@ -243,17 +220,17 @@ fi
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Start importing OpenStreetMap data: ${pbf_file} -> imposm3[./build/mapping.yaml] -> PostgreSQL"
echo "====> : Start importing OpenStreetMap data: ${area} -> imposm3[./build/mapping.yaml] -> PostgreSQL"
echo " : Imposm3 documentation: https://imposm.org/docs/imposm3/latest/index.html "
echo " : Thank you Omniscale! "
echo " : Source code: https://github.com/openmaptiles/openmaptiles-tools/tree/master/docker/import-osm "
echo " : The OpenstreetMap data license: https://www.openstreetmap.org/copyright (ODBL) "
echo " : Thank you OpenStreetMap Contributors ! "
make import-osm "PBF_FILE=${pbf_file}"
make import-osm
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Start importing border data from ${pbf_file} into PostgreSQL using osmborder"
echo "====> : Start importing border ${area} data into PostgreSQL using osmborder"
echo " : Source code: https://github.com/pnorman/osmborder"
echo " : Data license: http://www.openstreetmap.org/copyright"
echo " : Thank you: Paul Norman"
@ -288,7 +265,7 @@ echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : Start generating MBTiles (containing gzipped MVT PBF) from a TM2Source project. "
echo " : TM2Source project definitions : ./build/openmaptiles.tm2source/data.yml "
echo " : Output MBTiles: ./data/tiles.mbtiles "
echo " : Output MBTiles: ./data/${area}.mbtiles "
echo " : Source code: https://github.com/openmaptiles/openmaptiles-tools/tree/master/docker/generate-vectortiles "
echo " : We are using a lot of Mapbox Open Source tools! : https://github.com/mapbox "
echo " : Thank you https://www.mapbox.com !"
@ -314,11 +291,6 @@ rm -f ./data/quickstart_checklist.chk
cat ./data/quickstart_checklist.chk
ENDTIME=$(date +%s)
if stat --help >/dev/null 2>&1; then
MODDATE=$(stat -c %y "${pbf_file}" )
else
MODDATE=$(stat -f%Sm -t '%F %T %z' "${pbf_file}" )
fi
echo " "
echo " "
@ -333,15 +305,14 @@ docker images | grep openmaptiles
echo " "
echo "-------------------------------------------------------------------------------------"
echo "====> : (disk space) We have created the new vectortiles ( ./data/tiles.mbtiles ) "
echo "====> : (disk space) We have created the new vectortiles ( ./data/${area}.mbtiles ) "
echo " : Please respect the licenses (OdBL for OSM data) of the sources when distributing the MBTiles file."
echo " : Created from $pbf_file ( file moddate: $MODDATE ) "
echo " : Size: "
ls -la ./data/*.mbtiles
echo " : Data directory content:"
ls -la ./data
echo " "
echo "-------------------------------------------------------------------------------------"
echo "The ./quickstart.sh $osm_area is finished! "
echo "The ./quickstart.sh $area is finished! "
echo "It takes $((ENDTIME - STARTTIME)) seconds to complete"
echo "We saved the log file to $log_file ( for debugging ) You can compare with the travis log !"
echo " "