CHANGELOG.rst

Regolith Change Log

v0.5.1

Added:

• code to give more feedback to the user when the builder fails due to a database error

• regolith classlist can now read csv files in Columbia University format

• classlist register now checks whether a given file actually exists

• merge_collections to tools.py. merges two collections
• tests for manuscript review builder

• function for finding gaps and overlaps in lists of date-ranges
• utf8 support in all current latex builder templates

• function to dates that returns months as strings with leading zero where required

• function to dates that returns days as strings with leading zero where required

• ability to build publists with specified date ranges and filtered by grant

• tbd is now a valid month, returning 1 as an integer
• begin and end day now allowed in employment and education

Changed:

• unsegregated expense can now tolerate "tbd". This allows users to put a

  placeholder entry when the exact amount is not known, and then find it easily later. Code flags all tbd entries at build time, but doesn't crash.

• removed remote.rc logic from database.xsh
• current and pending builder extended to build c+p from merged proposal and grants collections

• added filter for cppflag so you can have current grants that don't appear in the current and pending form by setting cppflag to false in the db

• add needed_colls statement for quicker building
• moved has_started, has_finished and is_current to tools.py

• Load only dbs needed for builder, if builder declares which dbs it needs

• reimbursement builder requires a person to be specified on the command line

  to run due to the extreme slowness of openpyxl

• User supplied schemas now handles new keys in regolith validate.

Fixed:

• valueschema -> valuesrules in schema as valueschema deprecated in cerberus

• cpbuilder does name comparison on fuzzy-searched name for standardization

• cpbuilder includes initials when it is a multi-pi grant
• filter_grants in tools.py bug fixed that incorrectly reassigns team members
• updated docstring on filter_grants to make it clearer

• fix indenting of the editor eyes only block

• import from collections.abc not collections
• months can now be expressed as ints or strings as per the schema

• fix bug introduced in Jinja2 v2.11 that doesn't recognize conditional text

  in the import

• bug so that needed_colls results in only selected collections to be opened
• publist will now build even if person email and employment are missing

v0.5.0

Added:

• builders can now take --from and --to command-line args to specify date range
• added banner to groups schema, which is an image for website banner

None

• Google profile URL to people schema
• Research Focus Areas to people schema
• status to employment which will be selected from a list for sorting on the website

• filters in regolith.tools that return true if a given date is since or before or

  between other dates

• Add phone and address to CV and Resume if available

• builder for post-doc ad
• a builder for proposal reviews. Currently tuned for doe-bes and nsf-dmr
• builder for writing referee reports on manuscripts

• Make bib for entire group

• contacts to schema.py, a lighter type of person

Changed:

• builder now takes grant from grant field in expense and not by recursing

  into project

• if payee is direct_billed, builder will not build a reimbursement form

• all_documents now defaults to a deepcopy to prevent unintended mutation
• institutions schema to add street and make conditionals work better
• All months can now be integers or strings in the schemas
• Make a .bat file in scripts, which should help on windows
• now builds just accepted talks by default, not declined or pending

• proposals schema in schema.py to include fields for building current and

  pending report forms

• Use xonsh.lib.os.rmtree in conftest.py rather that building our own. The xonsh version is expected to do a better job on windows.

Removed:

None

• MTN: removed unused block from fuzzy_logic
• MTN: nicer handling of non-list objects in fuzzy_logic

Fixed:

• BUG: total amount now reproduces correctly in grants section

• BUG: account numbers not showing up in built reimbursement form

• Made the example current grant go to 2025 rather than 2018

• FIX: tests to run on windows OS by removing

  removed directory paths

• Makes sure some URLs in CV builder are also latex safe.
• correct spacing after date when it is a single day event
• Don't want to use latex_safe when we need the latex formatting
• Cast to string on way into latex_safe

• BUG: ints now handled the same as strings (appended) in fuzzy_logic

• BUG: now passes gtx as a list to fuzzy_logic not as a generator

v0.4.0

Added:

• Optional static_source key in the rc for the html build.

Changed:

• institution dereference is done by regolith.tools.dereference_institution function
• HTML pages dereference institutions
• person.html allows for authors or editors and hides publications in details
• root_index.html allows for banner to be speced in groups collection
• regolith.builders.CVBuilder now dereferences institutions/organizations for employers and education
• regolith.builders.CVBuilder deepcopies each person so we don't modify the records during dereference
• regolith.tools.latex_safe wraps URLs in \url{}
• regolith.builders.basebuilder.LatexBuilderBase runs pdflatex last if running on windows, rather than latex then dvipdf
• Order yaml collections by key before dump for deterministic changes in collection order (make git more sane)

Fixed:

• Properly handle authors and editors set in regolith.tools.filter_publications
• regolith.tools.fuzzy_retrieval properly handles null values
• education and employment subschemas for people are now just lists
• regolith.builders.BuilderBase uses latex_safe from regolith.tools
• wrap dbdir in @() so xonsh does the right thing

v0.3.1

Added:

• Schema for expenses tracking
• builder for Columbia reimbursement forms

Changed:

• open uses explict 'utf-8' bindings (for windows users)
• Allow education to be ongoing
• Allow begin and end years for service
• Make employment optional

Fixed:

• Build presentation PDFs when running in normal operation
• regolith.database.load_git_database checks branch gracefully
• regolith.tools.document_by_value doesn't splay address incorrectly

v0.3.0

Added:

• option for fuzzy_retrieval to be case insensitive
• regolith.broker.Broker for interfacing with dbs and stores from python
• regolith.builders.figurebuilder for including files from the store in tex documents
• regolith.database.open_dbs to open the databases without closing
• validate takes in optional --collection kwarg to restrict validation to a single collection
• ORCID ID in people schema
• Added presentations schema and exemplar
• Added institutions schema and exemplar
• Added presentation list builder
• number_suffix function to tools, returns the suffice to turn numbers into adjectives
• Method to find all group members from a given group
• a stylers.py module
• a function that puts strings into sentence case but preserving capitalization of text in braces
• User configuration file handling for adding keys to the regolithrc.json globally

Changed:

• added aka to groups schema
• Docs for collections fully auto generate (don't need to edit the index)
• zip and state only apply to USA institutions
• added group item in people schema
• KeyError for ChainDB now prints the offending key

None

• preslist now includes end-dates when meeting is longer than one day
• Builder for making presentation lists now builds lists for all group members
• Departments and schools in institutions are now dictionaries
• Preslist builder now puts titles in sentence case
• Use xonsh standard lib subprocess and os

Fixed:

• validate exits with error code 1 if there are bad records
• Preslist crash when institution had no department
• Departments and schools in institutions now use valueschema so they can have unknown keys but validated values

v0.2.0

Added:

• CPBuilder for building current and pending support reports
• initials field to people document
• person_months_academic, person_months_summer, and scope to grant document
• fuzzy_retrieval tool for getting documents based off of multiple potential fields (eg. name and aka for searching people)
• Tests for the exemplars
• Group collection for tracking research group information
• document_by_value tool for getting a document by it's value
• bibtexparser to test deps
• Builder integration tests
• Option for not making PDFs during the build process (for PDF building builders)
• Added presentations schema and exemplar
• Second exemplars for grants and proposals
• bootstrap_builders for generating the outputs to test the builders against
• publist tex file to tests

Changed:

• moved builders into builders folder
• group collection to groups collection
• Use the position sorter to enumerate the possible positions in the schema
• base.html and index.html for webpages are auto-generated (if not present)
• test against html in addition to other builders

Fixed:

• Pin to cerberus 1.1 in requirements. 1.2 causing testing problems.
• Fixed error that anded authors and editors
• Error in setup.py which caused builders to not be found
• Error in BaseBuilder which caused it to look in the wrong spot for templates
• Fixed bug in grad builder when the total wieght is zero.
• Actually use ChainedDB when working with the DBs
• Error in ChainedDB which caused bad keys to return with None

v0.1.11

Fixed:

• Local DBs were not being loaded properly

v0.1.10

Added:

• Regolith commands can run using a local db rather than a remote
• LatexBuilderBase a base class for building latex documents
• Users can override keys in each collection's schema via the RC
• Command for validating the combined database regolith validate

Changed:

• CVBuilder and ResumeBuilder builders now inheret from LatexBuilderBase

Fixed:

• Use get syntax with filter_publications in case author not in dict
• If a collection is not in the schema it is auto valid

v0.1.9

Fixed:

• all_documents now returns the values of an empty dict if the collection doesn't exist

v0.1.8

Added:

• Database clients now merge collections across databases so records across public and private databases can be put together. This is in client.chained_db.
• Blacklist for db files (eg. travis.yml) the default (if no blacklist is specified in the rc is to blacklist ['.travis.yml', '.travis.yaml']
• Schemas and exemplars for the collections. Database entries are checked against the schema, making sure that all the required fields are filled and the values are the same type(s) listed in the schema. The schema also includes descriptions of the data to be included. The exemplars are examples which have all the specified fields and are used to check the validation.
• Docs auto generate for collections (if they were documented in the schema).

Changed:

• all_docs_from_collection use the chained_db to pull from all dbs at once. This is a breaking API change for rc.client.all_documents
• App now validates incoming data against schema

Deprecated:

• Mongo database support is being deprecated (no chained_db support)

Fixed:

• Properly implemented the classlist replace operation.
• Fixed issue with classlist insertions using Mongo-style API (deprecated).
• Properly filter on course ids when emailing.
• fsclient dbs explicitly load 'utf-8' files, which fixes an issue on Windows

v0.1.7

Added:

• BuilderBase Class for builders
• Logo to docs
• Filesystem-based client may now read from YAML files, in addition to JSON. Each collection can be in either JSON or YAML.

Changed:

• Refactored builders to use base class

Fixed:

• Fixed issue with CV builder not filtering grants properly.
• Fixed bug with super not being called in the HTML builder.

v0.1.6

Added:

• Use Rever's whitespace parsing
• Fix template news

v0.1.5

Added:

• Rever release tool
• Interactive session support
• run better release

v0.1.4

Added:

• collabs field in db for collaborators
• active field in db for current collaborators/group members

Changed:

• People page only shows current members, former members on Former Members page