schemas/pg_hba.Mac.conf: Changed to match schemas/pg_hba.conf
schemas/pg_hba.conf: Fixed bug where also need an IPv6 bien entry with md5 authentication, because the IPv4 md5 authentication does not apply to "localhost", which is translated to the IPv6 address ::1
schemas/pg_hba.conf: Fixed bug where also need a local bien entry with md5 authentication, because the host-based md5 authentication applies only to literal IP addresses, not "localhost"
Added schemas/pg_hba.Mac.conf
schemas/pg_hba.conf: Restrict all accesses to the server except the bien user accessing vegbien using ident or a password, and the postgres superuser logging in using ident
inputs/.geoscrub/geoscrub_output/map.csv: Mapped countyvalidity to latLongInCounty
schemas/functions.sql: _map(): Fixed bug where entries that map to NULL were incorrectly being treated as if the entry didn't exist. Note that -> returns NULL both if the entry's value is NULL and if the entry doesn't exist, so ? must be used to recheck the presence of the key in the hstore.
mappings/VegCore-VegBIEN.csv: Mapped latLongInCounty
mappings/VegCore.csv: Added latLongInCounty
schemas/vegbien.sql: Added distance_to_county_m. Note that this can also be used to store latLongInCounty by mapping true to 0 and false to -1.
schemas/pg_hba.conf: Changed trust authentication back to ident/md5. Not sure how it got set to trust since I used md5 when enabling remote access to the DB for the bien user.
Added schemas/pg_hba.conf
schemas/vegbien.sql: place: Removed placecode to prevent datasources from creating duplicate entries for the same place, with different placecodes. This was a problem with the original BIEN2 geoscrub dataset, which contained duplicates.
inputs/import.stats.xls: Updated import times
Regenerated vegbien.ERD exports
schemas/vegbien.sql: analytical_stem_view: Fixed bug where need to join taxonoccurrence.collector_id to party because it's now an fkey rather than a literal name
schemas/vegbien.sql: analytical_*: Added coordinateUncertaintyInMeters
schemas/vegbien.sql: analytical_stem_view: Join to newWorldCountries on 2-digit ISO code instead of country name, to increase (BIEN2) newWorldCountries and GADM overlap
psql_vegbien: Run with sh because it no longer needs bash support
psql_script_vegbien: Fixed bug where needs to be run with bash instead of sh
Added inputs/newWorld/iso_code_gadm/
Added inputs/newWorld/_src/
inputs/XAL/Specimen/map.csv: darwin:FieldNumber: Removed command to determine that field is unused, because UNUSED is a factual assertion that does not need a reason to be specified each time
inputs/XAL/Specimen/map.csv: Remapped darwin:CoordinatePrecision to UNUSED
inputs/NY/Specimen/map.csv: Remapped CoordinatePrecision to coordinateUncertaintyInMeters, assuming units of m based on the range and precision of values
mappings/VegCore.csv: coordinatePrecision: Added units (degrees) to form coordinatePrecision_deg
mappings/VegCore-VegBIEN.csv: Removed mapping for coordinatePrecision, which is not the same as coordsaccuracy_m. coordinatePrecision is instead "the precision of the coordinates" themselves in degrees (<http://rs.tdwg.org/dwc/terms/#coordinatePrecision>).
schemas/vegbien.sql: coordinates: Changed coordinates.coordsaccuracy_deg units to m
Regenerated inputs/bien_web/observation/test.xml.ref
schemas/vegbien.ERD.mwb: Added projectcontributor, locationeventcontributor to ERD
schemas/vegbien.sql: higher_plant_group_nodes: Added root->NULL mapping to store all the families that don't match any higher plant group
schemas/vegbien.sql: higher_plant_group_nodes: Allow NULL values for higher_plant_group, to allow mapping all remaining families to NULL in family_higher_plant_group
psql_vegbien: Fixed bug where did not display command prompt when run from command line, by moving automatic setting of search_path to psql_script_vegbien. psql_script_vegbien is now used instead of psql_vegbien wherever the search_path needs to be set, so removing this functionality from psql_vegbien is not a problem.
input.Makefile: BIEN commands: $(psqlAsBien): Use psql_script_vegbien, which automatically adds the $(psqlOpts), instead of psql_vegbien
schemas/functions.sql: _map(): Support any entry having the value '*' (not just the '*' entry), which passes through that value. Support an entry having the value '!', which raises an exception.
inputs/SALVIAS/plotMetadata_/map.csv: AccessCode: Removed _map entry for 4, which does not apply to plots
schemas/vegbien.ERD.mwb: Fixed lines
schemas/vegbien.sql: locationevent: Added accesslevel
inputs/SALVIAS/plotMetadata_/map.csv: Mapped AccessCode to dcterms:accessRights with appropriate _map filter
Added inputs/.geoscrub/geoscrub_cleaned_unique/_no_import to disable geoscrub_cleaned_unique, since the new geoscrub_output supersedes it
Added inputs/.geoscrub/geoscrub_output/
Added inputs/.geoscrub/_src/README.TXT
Regenerated inputs/bien_web/observation/VegBIEN.csv
Added inputs/.geoscrub/_src/ to store Jim's geoscrub CSV
schemas/functions.sql: _map(): Changed error message for an unmapped value to "Value not in map" rather than "Invalid map value", because an unmapped value is not necessarily explicitly invalid
inputs/VegBank/plot_/map.csv: confidentialitystatus filter: Merged mappings for 0 with other public-equivalent fields. Note that fuzzed plots are still public, because the private columns have been removed.
inputs/VegBank/plot_/map.csv: Mapped confidentialitystatus to dcterms:accessRights with an appropriate _map filter
mappings/VegCore-VegBIEN.csv: Mapped dcterms:accessRights
schemas/functions.sql: _map(): Raise data_exception if value not in map and no default provided (not the same as a NULL default value)
mappings/VegCore-VegBIEN.csv: verbatimGrowthForm: Removed _map filter, which applied only to SALVIAS and has now been moved to the applicable SALVIAS tables
inputs/SALVIAS*/plotObservations/map.csv: Remapped Habit to growthForm with _map filter applied
sql_io.py: put_table(): Special handling for functions with hstore params: Fixed bug where need to unwrap literal values of mapping, which might be sql_gen.Literal objects
sql_gen.py: Added get_value()
dicts.py: join(): Added support for unhashable types, which are passed through. This adds support for SQL literal values which are dicts (hstores).
xml_func.py: Removed no longer used _map(), which has been replaced by a corresponding DB function
schemas/functions.sql: Added _map(), which uses the new hstore functionality. This expands _map() functionality to column-based import.
root Makefile: VegBIEN DB: DB and bien user: mk_db: hstore extension: Fixed bug where need to use `CREATE EXTENSION hstore SCHEMA pg_catalog` instead of createlang, because hstore must be explicitly created in pg_catalog or else it will be created in the public schema instead, causing it to get deleted every time the public schema is reinstalled and cascading the delete to everything (including in other schemas) that uses hstore
sql_io.py: put_table(): Added special handling for functions with hstore params. Note that although _map() doesn't exist yet as a DB function, this code must be in place before _map() is created to avoid param type mismatch errors.
root Makefile: PostgreSQL: postgres-Linux: Changed plpython to plpython3 in order to install plpython3u
schemas/py_functions.sql: _date(): Removed features that require dateutil, which is not available under plpython3u. This includes removing the now-unused date string parameter.
mappings/VegCore-VegBIEN.csv: Removed _date/date, because _date using a string date argument is no longer supported under plpython3u (dateutil is missing). Note that PostgreSQL's own date parsing is sufficient for most dates, so this use of _date is not strictly necessary and removing it will improve import times.
schemas/py_functions.sql: Replaced xrange() with range() for plpython3u
root Makefile: Python: python-Linux: Also install python3, needed by plpython3u
schemas/py_functions.sql: Updated except clause syntax for PostgreSQL 9.1.6
schemas/*.sql: Updated for PostgreSQL 9.1.6, which has standard_conforming_strings = on (which affects \-escapes in string literals), escape_string_warning not explicitly set, and uses ALTER TABLE ONLY instead of ALTER TABLE
README.TXT: Removed step to manually run make_analytical_db, now that this is done automatically by import_all. Added separate instructions to remake the analytical DB.
import_all: Change to main directory make targets are run from. Use relative paths to bin/ commands, which is possible now that the current dir is set.
import_all: Create a background process that waits until the import is done and then runs make_analytical_db
Added waitpid
import_all: Documented that `wait %1` waits for asynchronous commands
root Makefile: VegBIEN DB: DB and bien user: mk_db: Also install hstore extension. Note that this is only supported by PostgreSQL 9.1+.
input.Makefile: Editing import: Updated queries for current schema
inputs/.geoscrub/geoscrub_cultivated/create.sql: Fixed bug where need to filter out NULL lat/longs because primary keys can't contain NULL values
schemas/py_functions.sql: Changed function languages to plpython3u to match the new installed version. Note that plpythonu is not available on Mac under PostgreSQL 9.1.6.
reinstall_all: Fixed bug where also need to include datasources starting with . such as .TNRS/, by using with_all's new $all option
with_all: Added $all option to also include datasources starting with . such as .TNRS/. This is necessary for reinstall_all, which needs to install all datasources.
root Makefile: PostgreSQL: $(pg_ctl-*): Fixed bug where need to pause for a few seconds after restarting PostgreSQL, to wait for the server to be ready to accept connections
root Makefile: Installation: uninstall: Removed inputs/uninstall because the DB will be uninstalled anyway, so the inputs don't need to be individually removed first
schemas/postgresql.Mac.conf: Added back unix_socket_directory setting, which is apparently still needed in PostgreSQL 9.1.6
root Makefile: PostgreSQL: postgres-Linux: Also install postgresql.conf
root Makefile: PostgreSQL: postgres-Darwin: Also install postgresql.Mac.conf
root Makefile: PostgreSQL: $(macUsePostgresLib): Factored out PostgreSQL dir to $(macPostgresDir)
schemas/postgresql.Mac.conf: Updated to PostgreSQL 9.1.6's postgresql.conf
root Makefile: Datasources: inputs/install: Fixed bug where need to `wait` after `. bin/reinstall_all` to wait for inputs to finish installing before installing the public schema. This is necessary because views in the public schema now have dependencies on some datasources, such as TNRS.
root Makefile: PostgreSQL: $(psqlAsAdmin): Use new $(asAdmin)
root Makefile: VegBIEN DB: Schemas: schemas/public/install: Use $(psqlNoSearchPath) instead of $(psqlAsBien) because the search_path is set by vegbien.sql
root Makefile: Datasources: Added inputs/install override which runs `. bin/reinstall_all` instead, in order to install all datasources simultaneously
root Makefile: Python: python-Darwin: Added instructions to install Python 3.2 (Python 2 comes with Mac OS X, but Python 3.2 is needed for plpython3u)
root Makefile: VegBIEN DB: DB and bien user: mk_db: Updated for PostgreSQL 9.1.6 on the Mac, which only provides plpython3u (Python 3)
root Makefile: VegBIEN DB: DB and bien user: mk_db: Updated for PostgreSQL 9.1.6, which requires the DB name to be specified on the command line instead of in the $PGDATABASE env var set by postgres_vegbien. Fixed bug where need to run createlang as postgres superuser, because plpythonu is an untrusted language (with unrestricted access to the entire DB).
root Makefile: PostgreSQL: postgres-Darwin: Updated for PostgreSQL 9.1.6, which requires some /usr/lib/ symlinks to be changed to newer versions installed in the PostgreSQL lib/ dir
input.Makefile: $(psqlAsBien), csv2db: Turn off the automatic search_path where needed, because when the input is installed, the schemas in it may not exist yet
schemas/vegbien.sql: place: Renamed geosource_valid to geovalid. (It had gotten renamed in the reference -> source rename.)
schemas/vegbien.sql: location: Renamed confidentialitystatus->accesslevel, confidentialityreason->accessconditions to match the corresponding fields in source. Note that accessconditions stores more than confidentialityreason did, because it can contain details about the accesslevel in addition to the reason for it.
schemas/vegbien.sql: source.accesslevel, location.confidentialitystatus: Changed type to accesslevel
schemas/vegbien.sql: Added accesslevel enum
schemas/vegbien.sql: Renamed reference -> source to make this table more broadly applicable, and because this now stores the datasource metadata
schemas/vegbien.sql: referencename: Scope it by top-level datasource, because institutionCodes (which map to this field) are not globally unique. This involves renaming the previous reference_id field, which was for the matched reference, to matched_reference_id, to allow a scoping reference_id field.
