Texts might have several authors, in legacy_db this is expressed by columns tautor1, tautor2, tautor3
the default import script only covers tautor1, so customization is needed for the missing cols

same with places tort1, tort2, .. tort6

e.g. Text has CharFields start_date ... -> add additional field to store actual integers for start_end dates and provide some logic to populate them

provide GeoJson for SpatialCoverage including properties

• stelle
  • text.title
  • text.jahrhundert
  • text.start_date
  • text.end_date
    • text.author (is a M2M field ...)
    • text.art
• keyword
  • keyword.stichwort
  • keyword.wurzel
  • keyword.variants

the api already returns full object keyword arrays, it would be very helpful to have the possibility to display full objects for text.autor and text.ort

If a new Event object is created via web-application gui, make sure that the start_date field must not be empty on save
this can be definied by overriding/adapting archiv.forms.EventForm

see for example https://stackoverflow.com/a/47141693

• Autor.ort
• Stelle.text
• Stelle.key_word
• Text.autor
• Text.ort

add a table to the usecase detail view looking like

start-date | end-date | title | description | type [event/text/autor]

ordered by start-date (ascending) and presenting information for Events (all), related Authors and Texts -> #33 should help

Keywords should be filterable basically by all properties they are related to (through whatever relations), e.g. by Text, or the related place of the author of a text a stelle is related to
please add those additional filters

create a model method which returns an array with dicts looking like this

time_table_data = [
    {
        'id': 55,  # the object's id
        'start_date': 1104,  # the object's start_date
        'end_date': 1104,
        'ent_type': "event",  # lower case of the objects model name [event|autor|text]
        'ent_title': "1104 Lund replaces Hamburg as archbishopric of the North",  # title, or name
        'ent_description': "Lund replaces Hamburg as archbishopric of the North"  # only for events, for autor/text maybe some useful data or empty string
        'ent_detail_view': "/archiv/event/detail/55"  # the value of the .get_absolute_method
    },
    {
         'id': 234,  # the object's id
        ...
    }
]

implenting #31 before might be very helpful

currently theres no way to look for texts via keywords

potential new class "Event"
link to case-study?

Wäre es möglich, dass man Ergebnislisten auch als Excelliste herunterladen kann?

Excel ist möglich, allerdings braucht die Erstellung einer Excel Datei bei größeren Datenmengen ~ +1000 Zeilen extrem viel Resourcen; die Generierung einer .csv Datei wäre daher zu bevorzugen

Ort, Text, Stichwort, Autor

and add a simple CharField to class Text

to autor (gnd) and ort (geonames) as some LOD things are required

Edit und delete Button liegen nahe beieinander – könnte man das ‚delete‘ feld woanders unterbringen, um sicher zu gehen, dass nicht versehentlich etwas auf die Schnelle gelöscht wird?

adde choice options for this
options so far are 'Schlagwort' and 'Eigenname'

Neue Textstelle: kann man bei der Titeleingabe auch zwei Titel (oder mehrere) = eingeben; es gibt Texte (wenige), deren Titel nicht eindeutig auf einen einzigen eingrenzbar sind

with props

• title -> CharField
• description -> (Rich)TextField
• PI -> FK User

Passage M2M UseCase

part of FK -> KeyWord
related M2M -> KeyWord

parse https://gist.githubusercontent.com/csae8092/e51ce7f4a2d0bd5a9a2c162e96626aef/raw/f094562974de0df497a3b6a223c0694771655e97/events.txt and import at least the title and the description

running flake8 throws quit a lot of errors which should be fixed where possible
see also gh-action linting report https://github.com/acdh-oeaw/mmp/runs/2692857883

each archiv.Text object should be represented as XML/TEI object,

• including all related passages -> tei:body
• keywords -> tei:term
• Author(s) -> tei:header / tei:listPerson/tei:person
• bibliographic information -> tei:header

see for reference implementations

• gtrans
• apis_core

implementation (suggestion)

• render XML/TEI via django-template engine and parse the result with lxml to check if the created XML is actually well-formed

create convex-hull for spatial coverage + location of author;
probably as model-method for spatial coverage

override Event's save() method to make sure

1. if no end_date is provided, fill in the value of start_date
2. if end_date is smaller than start_date raise a custom error msg

create list/detail/edit/create/delete views + urls + templates for Event in analogy to the other classes

• icontains for title and description
• some greater than / lesser than for start_date / end_date

generate a graph based on cooccurrence of keywords in passages (stelle).

that is, e.g. from:

stelle1 | keyword1, keyword2
stelle2 | keyword1, keyword3
stelle3 | keyword1, keyword2, keyword3

generate:

keywordA | keywordB | count 
keyword1 | keyword2 | 2
keyword1 | keyword3 | 2
keyword2 | keyword3 | 1

additionally, we need to be able to filter by metadata of the stelle, especially: year, author, text

since we have data, we should use a subset of it to have fixtures
as data and datamodel changes, we need a way to automatically generate up to date fixtues

use https://github.com/davedash/django-fixture-magic

and write simple bash-script to run needed scripts

check if every list-view in app archiv actually returns a 200 http-code

current default filter takes tooooo much time to load

https://mmp.acdh-dev.oeaw.ac.at/archiv/stelle/ needs an author filter (autocomplete, multi select)

add a new view which exposes the timetable data #33 of a given UseCase as JSON
the url should look like:
usecase-timetable-data/(?P<pk>[0-9]+)$ so e.g. http://mmp.acdh-dev.oeaw.ac.at/archiv/usecase-timetable-data/7

• use from django.http import JsonResponse see e.g. https://simpleisbetterthancomplex.com/tutorial/2016/07/27/how-to-return-json-encoded-response.html
• if a given primary key does not match any UseCase, raise a 404 Error (see https://docs.djangoproject.com/en/3.2/topics/http/shortcuts/#django.shortcuts.get_object_or_404)

Suchfunktion: wichtig wäre es Autor und key word gemeinsam suchen zu können; d.h. keyword in den Werken eines Autors; um u.a. auch sehen zu können wie oft ein Autor ein bestimmtes keyword benutzt und in Verbindung mit welchen anderen keywords

actually this would be nice not only for dates but also for CharField Filter, so that users can choose if the lookup is exact, iexact, or e.g. startswith

these markup leads to errors when parsed with lxml etree

rewrite the usecase detail view template in a way that the table for the timetable is not rendered by django-template but with vanilla javascript

1. make an ajax request to the timetable endpoint (#36)
2. process the response to create a table presenting the data in the html

and yes, thats much more complicated than the already existing solution, but its a nice practice, and this technique might become useful if we are dealing with big amounts of data

add those fields to all classes

in DetailView of Stelle, click on button "create Spatialcoverage" -> links to CreateView of Stelle with 'stelle' already populated

Stelle.zitat contains the actual quote (i.e. the text transcribed/copied from the source) AND the location, i.e. p. 123-344. good thing though, it looks like the actual quote starts after the first :, so it should be easy move this information into a dedicated field

-> this could be added to the import_data script

Move the JavaScript code into a separate file in archiv/static/archiv/js/tt-fetch.js and just call the actual function in the use-case detail view template

provide methods for archiv.models.UseCase to fetch Text ,Autor and related KeyWord which are somehow (via Stelle) related to the given UseCase.

use the cached-property method decorator so that the related objects can be called as any other properties like some_usecase.title

see for a reference implementation https://github.com/acdh-oeaw/mmp/blob/master/archiv/models.py#L544

also read about cached-properties in the django-docs
also I assume that some clever ordering does matter, e.g. first fetch all related 'Stelle' objects and use those in the methods to fetch Texts and use this method/property for fetching Autors