added inputs/VegBank/verify/input_cols.txt, input_cols.unmapped.txt (with runscript to filter input_cols.unmapped.txt)
schemas/VegBIEN/attribution/BIEN 3 data use and attribution.docx: made Ramona's corrections with track changes turned on. note that you have to use MS Word for this, not LibreOffice, because LibreOffice can't save the table of contents properly in .docx or .doc format (although it can save it in .odt format).
inputs/VegBank/stratum/postprocess.sql: added pkey
inputs/VegBank/taxonobservation_/postprocess.sql: added __parent index on locationID to facilitate the LEFT JOINs used to create the validation input
inputs/VegBank/observation_/postprocess.sql: added __parent index on locationID to facilitate the LEFT JOINs used to create the validation input
inputs/VegBank/import_order.txt: added taxon_observation.**.sample so it will automatically be kept up to date
inputs/VegBank/taxon_observation.**.sample/create.sql: set runtime (1 s)
inputs/VegBank/: added taxon_observation.**.sample subset of plots to use in the validation. this avoids the need to import all of VegBank just to validate a few of the plots.
bin/map: usage: documented that verbosity > 3 in commit mode turns on debug_temp mode, which creates real tables instead of temp tables
schemas/VegCore/ERD/VegCore.ERD.mwb: georeferencing: documented that this stores location determinations (= VegBIEN.vegpath.org?public.locationplace)
inputs/VegBank/taxon_observation.**/: updated for data refresh
inputs/VegBank/plantconcept_/: mapped columns, since this is now included in import_order.txt and therefore gets processed by the column-renaming runscripts. note that this means that in taxonobservation_/map.csv, the plantconcept_ input column names need to be changed to what they are mapped to.
inputs/VegBank/taxonobservation_/create.sql: updated runtime (20 s)
schemas/VegCore/ERD/VegCore.ERD.mwb: regenerated exports and udpated image map
fix: schemas/VegCore/ERD/VegCore.ERD.mwb: project: allow it to have a dataset rather than being one, because projects do not always have separate attribution requirements (e.g. in Madidi they don't). this also fixes a bug where projects whose datasets had a parent dataset (e.g. VegBank) would need a dummy event for that parent dataset so that the fkeys on both dataset.parent->dataset and event.parent->event would be satisfied. added project__unique constraint on name within dataset.
schemas/VegCore/ERD/VegCore.ERD.mwb: taxa_sampling_event.parent (inherited): require this to be a taxa_sampling_event. documented that this is the taxa_sampling_event for the parent stratum.
schemas/VegCore/ERD/VegCore.ERD.mwb: fixed lines
schemas/VegCore/ERD/VegCore.ERD.mwb: place_visit.parent (inherited): added fkey constraint that this must be a place_visit. documented that this is the concurrent observation of the parent place.
schemas/VegCore/ERD/VegCore.ERD.mwb: project: inherit from dataset, since projects generally confer attribution and conditions of use
fix: schemas/VegCore/ERD/VegCore.ERD.mwb: place_visit: fixed synonyms list to "VegX.vegpath.org?plotObservation, VegBank.vegpath.org?observation (which was confusingly named) + some of VegBank.vegpath.org?plot (which is actually 1:1 with observation)"
schemas/VegCore/ERD/VegCore.ERD.mwb: place_observation: renamed to place_visit for clarity, since other tables that point to this use it as a general container, not just a link to the observations about the place
schemas/VegCore/ERD/VegCore.ERD.mwb: merged stratum_event into taxa_sampling_event since they serve the same purpose (separating taxa samplings within the same place_observation by subsetting methodology). merged layer into stratum since strata as we think of them are horizontal, although there are other definitions of stratum (e.g. size class) which are more similar to our method table.
schemas/VegCore/ERD/VegCore.ERD.mwb: event subclasses: subject: added applicable fkey constraints to the inherited field. renamed existing table-specific subject fields to subject to merge with the inherited field.
schemas/VegCore/ERD/VegCore.ERD.mwb: event: added unique constraints for every way that the scoping fields can meaningfully be combined. these correspond to the various unique constraints offered by VegBIEN.locationevent. (the event table's constraints are always tricky because there are many ways to define an event. it is important to include all of them so that there is almost always a way to generate a natural key, without resorting to an artificial key from traceable.id_by_source.)
schemas/VegCore/ERD/VegCore.ERD.mwb: event.place: renamed to subject and allow this to be any traceable, not just a place. this allows taxon_observations to be scoped within their taxon_occurrences just as subplot events are scoped within parent plot events.
schemas/VegCore/ERD/VegCore.ERD.mwb: traceable.authors: documented that this makes traceable mutually recursive with party_list
schemas/VegCore/ERD/VegCore.ERD.mwb: fixed inheritance connectors
schemas/VegCore/ERD/VegCore.ERD.mwb: record.authors: moved to traceable because actually, different sections of a denormalized source record may have separate authors (e.g. participants for taxa_sampling_event, identifiers for taxon_determination)
schemas/VegCore/ERD/VegCore.ERD.mwb: record.dataset: split this into the scoping_dataset (which scopes the id_within_dataset) and the attribution_dataset (which provides attribution information)
schemas/VegCore/ERD/VegCore.ERD.mwb: record.authors: changed this to refer only to original data creators such as collectors and identifiers. this way, there is no need to generate a combined party_list of owners+authors for every record, because the record-specific authors list can just point to the same party_list as the collectors or event participants. this is also closer to Brad's meaning, which is specifically "a person or institution responsible for creating the dataset" [+emph].
schemas/VegCore/ERD/VegCore.ERD.mwb: party_list: added array field, which stores an array combining all the list entries. if specified, a trigger will autogenerate party_list_entries from this field. if not specified, it will be autopopulated whenever a party_list_entry is added.
schemas/VegCore/ERD/VegCore.ERD.mwb: added data_owner subclass of party_list_entry, which stores use_conditions for each data_owner of a dataset. dataset: added access_rights (public, private, etc.), use_conditions (citation requirements, etc.).
schemas/VegCore/ERD/VegCore.ERD.mwb.run: record.authors: clarified that this should include all parties who helped create the data record, making it a superset of dataset.data_owners, who receive attribution. clarified that record-specific authors such as collectors and identifiers generally do not receive attribution, although dataset-level authors do (this field includes both).
schemas/VegCore/ERD/VegCore.ERD.mwb.run: record.id_within_dataset: removed instruction to also put this in source.name, because source.name is now optional, and this would not make sense because source does not have a parent pointer
schemas/VegCore/ERD/VegCore.ERD.mwb.run: documented that schema COMMENTs cannot contain ' (used in contractions such as "don't"), because MySQL Workbench cannot escape these properly
schemas/VegCore/ERD/VegCore.ERD.mwb: record: added authors, which stores the people who helped create the data record. this is Brad's data author term (Brad.vegpath.org?attribution.data_author), but actually belongs to the individual data record rather than to an entire data*set* because records may have different authors (collectors, identifiers, etc.). note that schema COMMENTs apparently cannot contain ' (used in contractions such as "don't"), because MySQL Workbench cannot escape these properly.
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset.data_owners: documented that this includes all aggregators through which the data has passed, so that this is equivalent to Brad's data indexers (Brad.vegpath.org?attribution.data_indexer)
fix: schemas/VegCore/ERD/VegCore.ERD.mwb: dataset.parent: fixed example to indicate that the GBIF.ARIZ dataset is contained within a specific GBIF dataset (the parent), but was obtained from the GBIF organization (the dataset_source). ARIZ is then the first_publisher, not the dataset_source.
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset: renamed data_source to dataset_source for clarity
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset.contacts: documented that this is the contact person(s) to ask about attribution questions. there should always be at least one contact person. if unknown, the person loading the data should set this to themselves.
added schemas/VegCore/Brad_Boyle/BIEN 3 data use and attribution.docx symlink to schemas/VegBIEN/attribution/BIEN 3 data use and attribution.docx, so that all of Brad's attribution-related terms are in one place
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset.contacts: make this required, so that anyone adding data to a VegCore DB always provides at least one contact person, possibly themselves, to ask about attribution questions (as suggested by Brad in BIEN 3 data use and attribution.docx)
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset.first_publisher: allow this to be any kind of source, rather than just a party (i.e. an organization). this is possible now that party inherits from source.
schemas/VegCore/ERD/VegCore.ERD.mwb: dataset: linked to Brad's synonyms for dataset_source (Brad.vegpath.org?attribution.proximate_data_provider), first_publisher (Brad.vegpath.org?attribution.data_publisher, Brad.vegpath.org?data_provenance.primarydataprovider)
bugfix: web/people/Brad_Boyle/.htaccess: only forward requests for a term defined by Brad to the attribution/ subdir, so that Brad.vegpath.org shows a directory listing instead of forwarding right to the attribution doc
bugfix: web/people/Brad_Boyle/.htaccess: set %{ENV:innermost_dir} to Brad_Boyle
added web/people/Brad_Boyle/data_provenance/ (BIEN database entities related to data provenance and ownership.docx)
bugfix: web/people/Brad_Boyle/attribution/.htaccess: \-escape the %s so that they aren't treated as backreferences. set the [redirect] flag so that the downloaded file has the correct filename (BIEN 3 data use and attribution.docx instead of attribution).
bugfix: web/people/Brad_Boyle/attribution/.htaccess: use %20 instead of + because the Apache server apparently doesn't translate + to space correctly
bugfix: web/people/Brad_Boyle/attribution/.htaccess: link to the VegBIEN subdir in /exchange_schemas instead of the main /VegBIEN because there are actually two VegBIEN subdirs, and the main one instead points to /datasources/VegBIEN
added web/people/Brad_Boyle/ with attribution/ subdir for BIEN 3 data use and attribution.docx
inputs/VegBank/plantconcept_/create.sql: documented runtime (21 min)
bugfix: inputs/VegBank/plantconcept_/: added new-style import files
bugfix: inputs/VegBank/import_order.txt: added plantconcept_, because new-style import needs it to be explicitly listed in import_order.txt in order to run it
inputs/VegBank/run: refresh(): added usage
inputs/VegBank/run: refresh(): documented that this should be run on vegbiendev
inputs/VegBank/_archive/2012-8-30/: svn:ignore the data exports
inputs/VegBank/run: added refresh() target
inputs/VegBank/: refreshed VegBank so that all of Mike Lee's sample plots would be included in the input data. (VegBank was last refreshed from the live DB on 2012-8-30.) split vegbank.sql into vegbank.schema.sql and vegbank.data.sql so that the schema can be examined and imported separately, like for MySQL datasources. inputs/VegBank/vegbank.~.clean_up.sql: commented out setting comminterpretation.commname to NOT NULL, because after the refresh it is now NULL in 10 rows, where commconcept_id is also NULL.
lib/runscripts/datasrc_dir.run: import(): added remake (rm=1) mode that reinstalls the datasource before continuing with the subdirs' import actions
lib/runscripts/in_datasrc_dir.run: added datasrc_make(), which runs make in the datasrc dir
inputs/.geoscrub/_src/README.TXT: added e-mail from John Donoghue with general description of the BIEN2 geovalidation workflow
bugfix: schemas/vegbien.sql: plot.**: removed ORDER BY because a SELECT on the view with its own ORDER BY will actually sort the rows twice instead of optimizing away plot.**'s default ORDER BY
schemas/vegbien.sql: added plot.** subset view of analytical_stem_view, used for VegBank validation
planning/timeline/timeline.2013.xls: moved reload core & analytical database out a week because it shouldn't be done until after CVS validation is complete
inputs/.geoscrub/_src/README.TXT: added link to geovalidation description in wiki
added schemas/VegBIEN/attribution/BIEN 3 data use and attribution.docx from Brad
planning/timeline/timeline.2013.xls: updated to reflect that the next week should be spent doing data validations, as decided in the conference call
bugfix: schemas/vegbien.sql: locationevent__communities(): need to use commdescription instead of commname as the communityName because that's where communityName has been mapped to in mappings/VegCore-VegBIEN.csv. (commname instead stores the communityID.)
inputs/VegBank/: added observation__community/
inputs/VegBank/vegbank.~.clean_up.sql: commclass.commcode,commname: rename to prevent collisions
inputs/VegBank/vegbank.~.clean_up.sql: indicate required column comminterpretation.commname
inputs/VegBank/vegbank.~.clean_up.sql: commconcept.commname: rename to prevent collision with commname.commname
planning/timeline/: svn:ignore: added .~* (lock files for .xls)
schemas/vegbien.sql: analytical_stem_view: added instructions for what to do after updating this
schemas/vegbien.sql: analytical_stem_view, analytical_plot: added communities using new locationevent__communities__array()
schemas/vegbien.sql: added locationevent__communities*()
bugfix: schemas/vegbien.sql: locationevent: locationevent__pull_forward_from_parent() trigger: only pull forward the location_id when there is a stratum_id, so that empty locationevents that should be auto-forwarded are not unintentionally made valid by populating the location_id field (which causes a check constraint to become satisfied, and prevents auto-forwarding from working). this fixes a bug in some datasources' automated tests, such as `make inputs/SALVIAS/plotObservations/test`.
lib/sql_io.py: put_table(): default param: documented that this will be used for all missing rows, regardless of which error caused them not to be inserted. this means that auto-forwarding (wiki.vegpath.org/Auto-forwarding) can be used with any type of constraint violation, not just NOT NULL constraints (which it is typically used with).
/Makefile: mk_db: config/users.sql: unset ON_ERROR_STOP so user-exists errors don't prevent further user-adds
planning/timeline/timeline.2013.xls: updated for progress
bugfix: schemas/vegbien.sql: analytical_stem_view: plot__collectors: to include plot__collectors for nested stratum events, need to calculate the locationevent__contributors__array() for the locationevent and parent_event separately, and then COALESCE those, instead of just running locationevent__contributors__array() once on the COALESCE of the two IDs
bugfix: schemas/vegbien.sql: locationevent: added pull-forward trigger to populate location_id from the parent event. a pull-forward trigger must be used for this (not a COALESCE in analytical_stem_view) because the locationevent cannot even be found by analytical_stem_view if it does not have an associated location.
schemas/VegCore/ERD/VegCore.ERD.mwb: source: added required url field, which points to the source data and uniquely identifies the source. note that this differs from the id, which is a self-refential pkey (as described at wiki.vegpath.org/Web_interface#Clickable-fkeys).
schemas/VegCore/ERD/VegCore.ERD.mwb: source: removed canon, because this would create too much complexity as a result of having multiple copies of the same source with different names. instead, source names will be scrubbed upon insert so that the row always contains the most accurate data, and there is only one copy of each source. this scrubbing may use the assistance of a thesaurus table w/ standard abbreviations, etc.
schemas/VegCore/ERD/VegCore.ERD.mwb: plot.name: merged with inherited name field
schemas/VegCore/ERD/VegCore.ERD.mwb: source: added canon field. when there are multiple names for a source, this points to the canonically-named source. by default, points recursively to itself.
schemas/VegCore/ERD/VegCore.ERD.mwb: taxon_observation.taxon_occurrence: documented that when not provided (and not fillable from another field), a taxon_occurrence will be created with the same id as the taxon_observation. this operation should eventually be perfomed by a trigger.
schemas/VegCore/ERD/VegCore.ERD.mwb: geovalidatable_place.parent_geoplace: renamed to parent so it will merge with the inherited parent field. this is the parent geoplace.
