Regolith

====================

**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

====================

**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

====================

**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

====================

**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

====================

**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``

====================

**Fixed:**

* Local DBs were not being loaded properly