input.Makefile: Staging tables: import/install-%: Only install staging table if input contains only CSV sources. Changed $(isXml) to $(isCsv) (negated) everywhere because rules almost always only run something if input contains only CSV sources, rather than if input contains XML sources.
input.Makefile: Staging tables: import/install-%: Output load status to log file if log option is set
Scripts that are meant to be run in the calling shell: Fixed bug where running the script inside another script would make the script think it was being run as a program, and abort with a usage error
Scripts that are meant to be run in the calling shell: Fixed bug where running the script as a program (without initial ".") wouldn't be able to call return in something that was not a function. Converted all code to a <script_name>_main method so that return would work properly again. Converted all variables to local variables.
env_password: return instead of exit if password not yet stored, in case user is running it from a shell without the initial "-" argument. (This would be the case if the user is just testing out the script, instead of using a command that env_password directs them to run.)
env_password: Use ${BASH_SOURCE0} for $self and $self for $0. return instead of exit on usage error in case user is running it from a shell.
stop_imports: Use ${BASH_SOURCE0} for $self and $self for $0
import_all: Use new with_all. Use ${BASH_SOURCE0} for $self and $self for $0.
Added with_all to run a make target on all inputs at once
Made row #s 1-based to the user to match up with the staging table row #s
bin/map: Fixed bug where limit passed to sql.select() was end instead of the # rows, causing extra rows to be fetched when start > 0. Documented that row #s start with 0.
Removed no longer needed csv2ddl
input.Makefile: Staging tables: import/install-%: Use new csv2db instead of csv2ddl/$(psqlAsBien), because it handles translating encodings properly
Added csv2db to load a command's CSV output stream into a PostgreSQL table
schemas/postgresql.Mac.conf: Set unix_socket_directory to the appropriate Mac OS X dir, since otherwise, the socket is apparently not created and `make reinstall_db` doesn't work
main Makefile: VegBIEN DB: db: Set LC_COLLATE and LC_CTYPE explicitly, to make it easier to change them
Added ProgressInputStream
exc.py: print_ex(): Added plain option to leave out traceback
main Makefile: VegBIEN DB: db: Use template0 to allow encodings other than UTF-8. Because template0 doesn't have plpgsql on PostgreSQL before 9.x, add "CREATE PROCEDURAL LANGUAGE plpgsql;" manually in schemas/vegbien.sql.make, and filter it back out on PostgreSQL after 9.x using db_dump_localize.
PostgreSQL-MySQL.csv: Remove "CREATE PROCEDURAL LANGUAGE" statements
Added db_dump_localize to translate a PostgreSQL DB dump for the local server's version
vegbien_dest: Added option to override the prefix of the created vars
schemas/vegbien.sql.make: Fixed bug where data sources' schemas were also exported by exporting only the public schema. Note that this also removes the "CREATE OR REPLACE PROCEDURAL LANGUAGE plpgsql" statement, so that it doesn't have to be filtered out with `grep -v`.
input.Makefile: input.Makefile: Use `$(catSrcs)|` instead of $(withCatSrcs) where possible
sql.py: pkey(): Fixed bug where results were not being cached because the rows hadn't been explicitly fetched, by having DbConn.DbCursor.execute() fetch all rows if the rowcount is 0 and it's not an insert statement. DbConn.DbCursor: Made _is_insert an attribute rather than a method, which is set as soon as the query is known. Added consume_rows(). Moved Result retrieval section above Database connections because it's used by DbConn.
sql.py: pkey(): Fixed bug where queries were not being cached. Use select() instead of run_query() so that caching is automatically turned on and table names are automatically escaped.
streams.py: Added LineCountInputStream, which is faster than LineCountStream for input streams. Added InputStreamsOnlyException and raise it in all *InputStream classes' write() methods.
sql.py: DbConn: For non-cacheable queries, use a plain cursor() instead of a DbCursor to avoid the overhead of saving the result and wrapping the cursor
Moved db_config_names from bin/map to sql.py so it can be used by other scripts as well
csv2ddl: Also print a COPY FROM statement
input.Makefile: Fixed bug where input type was considered to be different things if both $(inputFiles) and $(dbExport) are non-empty. Now, $(inputFiles) takes precedence so that the presence of any input files will cause a DB dump to be ignored. This ensures that a (slower) input DB is not used over a (faster) flat file.
csvs.py: stream_info(): Added parse_header option. reader_and_header(): Use stream_info()'s new parse_header option.
csv2ddl: Renamed schema name env var from datasrc to schema to reflect what it is, and to make the script general beyond importing inputs
input.Makefile: Moved Installation, Staging tables after Existing maps discovery because they depend on it. Staging tables: Create a staging table for each table a map spreadsheet is available for. Put double quotes around the schema name so its case is preserved.
Added csv2ddl to make a PostgreSQL CREATE TABLE statement from a CSV header
sql.py: Input validation: Moved section after Database connections because some of its functions require a connection. Added esc_name_by_module() and esc_name_by_engine(), and use esc_name_by_module() in esc_name().
input.Makefile: Installation: Create a schema for the datasource in VegBIEN as part of the installation process. This will be used to hold staging tables.
input.Makefile: Changed install, uninstall to depend on src/install, src/uninstall targets, which in turn depend on db, rm_db. This will allow us to add additional install actions for all input types.
sql.py: DbConn: Cache the constructed CacheCursor itself, rather than the dict that's used to create it
sql.py: pkey(): Changed to use the connection-wide caching mechanism rather than its own custom cache. DbConn.__getstate__(): Don't pickle the debug callback.
sql.py: DbConn: Added is_cached(). run_query(): Use new DbConn.is_cached() to avoid creating a savepoint if the query is cached.
sql.py: DbConn: Also cache cursor.description
sql.py: DbConn: Cache query results as a dict subset of the cursor's key attributes, so that additional attributes can easily be cached by adding them to the subset list
dicts.py: Added AttrsDictView
util.py: NamedTuple.__iter__(): Removed unnecessary **attrs param
sql.py: _query_lookup(): Fixed bug where params was cast to a tuple, even though it could also be a dict. index_cols(): Changed to use the connection-wide caching mechanism rather than its own custom cache.
util.py: NamedTuple: Made it usable as a hashable dict (with string keys) by adding iter() and getitem()
dicts.py: Added make_hashable()
sql.py: DbConn: Only cache exceptions for inserts since they are not idempotent, but an invalid insert will always be invalid. If a cached result in an exception, re-raise it in a separate method other than the constructor to ensure that the cursor object is still created, and that its query instance var is set.
sql.py: insert(): Cache insert queries by default. This works because any DuplicateKeyException, etc. would be cached as well. This saves many inserts for rows that we already know are in the database.
sql.py: DbConn.run_query(): Cache exceptions raised by queries as well
sql.py: DbConn.run_query(): When debug logging, label queries with their cache status (hit/miss/non-cacheable)
sql.py: DbConn.run_query(): Also debug-log queries that produce exceptions
sql.py: DbConn: Allow creator to provide a log function to call on debug messages, instead of using stderr directly
bin/map: Pass debug mode to DbConn so that SQL query debugging works again
sql.py: DbConn: DbCursor: Fixed bug where caching was always turned on, by passing the cacheable setting to it from run_query(). Turned caching back on (uncommented it) since it's now working.
bin/map: map_rows()/map_table(): Pass kw_args to process_rows() so rows_start can be specified when using them. DB inputs: Skip the pre-start rows in the SQL query itself, so that they don't need to be iterated over by the cursor in the main loop.
bin/map: Fixed bug introduced in r1718 where the row # would not be incremented if i < start, causing an semi-infinite loop that only ended when the input rows were exhausted. process_rows(): Added optional rows_start parameter to use if the input rows already have the pre-start rows skipped.
input.Makefile: Sources: cat: Changed Usage message to use "--silent" make option
input.Makefile: Sources: cat: Added Usage message with instructions for removing echoed make commands
run_*query(): Fixed bug where INSERTs, etc. were cached by making callers (such as select()) explicitly turn on caching. DbConn.run_query(): Fixed bug where cur.mogrify() was not supported under MySQL by making the cache key a tuple of the unmogrified query and its params instead of the mogrified string query. CacheCursor: Store attributes of the original cursor that we use, such as query and rowcount.
sql.py: Made row() and value() cache the result by fetching all rows before returning the first row
iters.py: Added func_iter() and consume_iter()
sql.py: Cache the results of queries (when all rows are read)
Proxy.py: Fixed infinite recursion bug by removing setattr() (which prevents the class and subclasses from storing instance variables using "self." syntax)
sql.py: DbConn: Added run_query(). run_raw_query(): Use new DbConn.run_query().
Added Proxy.py
parallel.py: MultiProducerPool: Added code to create a shared Namespace object, commented out. Updated share() doc comment to reflect that it will writably share the values as well.
bin/map: Share locals() with the pool at various times to try to get as many unpicklable values into the shared vars as possible
dicts.py: Turned id_dict() factory function into IdDict class. parallel.py: MultiProducerPool: Added share_vars(). main_loop(): Only consider the program to be done if the queue is empty and there are no running tasks.
collection.py: rmap(): Treat only built-in sequences specially instead of iterables. Pass whether the value is a leaf to the func. Added option to only recurse up to a certain # of levels.
Added lists.py
collection.py: rmap(): Fixed bugs: Made it recursive. Use iters.is_iterable() instead of isinstance(value, list) to work on all iterables. Use value and not nonexistent var list_.
iters.py: Added is_iterable()
parallel.py: prepickle(): Pickle all objects in vars_id_dict_ by ID, not just unpicklable ones. This ensures that a DB connection created in the main process will be shared with subprocesses by reference (id()) instead of by value, so that each process can take advantage of e.g. shared caches in the connection object. Note that this may require some synchronization.
parallel.py: MultiProducerPool.main_loop(): Got rid of no longer correct doc comment
bin/map: Share on_error with the pool
parallel.py: MultiProducerPool: Pickle objects by ID if they're accessible to the main_loop process. This should allow e.g. DB connections and pools to be pickled, if they were defined in the main process.
Added dicts.py with id_dict() and MergeDict
Added collection.py with rmap()
db_xml.py: put(): Moved pool.apply_async() from put_child() to put_(), and don't use lambdas because they can't be pickled
parallel.py: MultiProducerPool.apply_async(): Prepickle all function args. Try pickling the args before the queue pickles them, to get better debugging output.
sql.py: with_savepoint(): Use new rand.rand_int()
rand.py: rand_int() Fixed bug where newly-created objects did not have unique IDs because they were on the stack. So, we have to use random.randint() anyway.
Added rand.py
sql.py: DbConn: Made it picklable by establishing a connection on demand
bin/map: Also consume asynchronous tasks before closing the DB connection (this is where most if not all tasks will be consumed)
Runnable.py: Made it picklable
Added eval_.py
Added Runnable
db_xml.py: put(): Added parallel processing support for inserting children with fkeys to parent asynchronously
parallel.py: Fixed bugs: Added self param to instance methods and inner classes where needed
parallel.py: Changed to use multi-producer pool, which requires calling pool.main_loop()
parallel.py: Pool: Added doc comment
parallel.py: Pool: apply_async(): Return a result object like multiprocessing.Pool.apply_async()
bin/map: Use new parallel.py for parallel processing
Added parallel.py for parallel processing
bin/map: Use dummy synchronous Pool implementation if not using parallel processing
bin/map: Use multiprocessing instead of pp for parallel processing because it's easier to use (it uses the Python threading API and doesn't require providing all the functions a task calls). Allow the user to set the cpus option to to use all system CPUs (needed because in test mode, the default is 0 CPUs to turn off parallel processing).
