xmlfiles

xml-files

• Lexicon files in crk/src/
  • N_crk.xml, V_crk.xml, Ipc_crk.xml, MWE_crk.xml
• Lexicon files in crk/engcrk/
  • N_engcrk.xml, V_engcrk.xml, Ipc_crk.xml, MWE_crk.xml
• Lexicon files in crk/fracrk/
  • N_fracrk.xml, V_fracrk.xml, Ipc_crk.xml, MWE_crk.xml
• Put the semantic sets into semantic fields: meta/semantic_sets.xml
• MorfaC questions files (in meta):
  • verb_questions.xml
  • noun_questions.xml
• Feedback/hints (in meta):
  • feedback_nouns.xml, feedback_verbs.xml, messages.eng.xml
• Other files for MorfaC (in meta):
  • fillings.xml, grammar_defaults.xml

Lexicon files

Elements are

• <r> root (the whole file)
• <e> entry, one for each crk lemma
• <lg> lemma group, contains one lemma and one sources
• <l> lemma, with attributes: pos (part of speech), animacy for nouns (IN, AN) or trans_anim for verbs (AI, II, TI, TA), audio (if there is an audiofile for the lemma in crk_oahpa/media/audio/), gen_only for restricting the generation. gen_only="none" gives no generation at all, the lemma will only be used for Leksa. The attribute object for giving context to the verbs, showing that they are transitive, and if they take animate or inanimate noun.
• <sources>
• <book> with attributes: name (book chapter or ). The book chapters are put together in levels ane added to the menu in these files: crk_oahpa/crk_drill/game.py and crk_oahpa/crk_drill/forms.py
• <mg> meaning group
• <semantics>
• <sem> with attributes: class (several). Semantic classes starting with an m in the name (mACTIVITY) is used only in Morfa. One can use any name for the semantic classes, but they have to be added to this file: meta/semantic_sets.xml, to be a part of the semantic choices in the menu of Leksa. E.g. ANIMAL can be a part of the semantic field NATURE.
• <tg> translation group with attributes xml: lang= for language
• <t> translation with attributes: pos (part of speech), stat="pref" for the preferred translation to be shown as key answer

Question files for Morfa C

Explanations about how this works and how to make the tasks

• <questions> (the whole file)
• <q> with attributes: id (a uniq id for each task, e.g. npl1: noun plural 1
• <qtype> type of task: for the menu
• <question> - one in each <q> element
• <text> - the question with variables=elements
• <element>: definition of one element, with attribute id=
• <grammar> with attribute tag, to be matched with the string in the data base (output from FST). One can use variables, which have to be defined here: grammar_defaults.xml
• <sem> with attribute class, for restricting the choice of lemma
• <answer>. There can be more than one possible answer to the question
• <text> - the answer with variables=elements
• <element>: definition of one element, with attribute id= and task="yes" for the slot, game="morfa"
• <grammar> with attribute tag, to be matched with the string in the data base (output from FST). One can use variables, which have to be defined here: grammar_defaults.xml
• <sem> with attribute class, for restricting the choice of lemma

Generating forms

Files for generation forms are in meta: n_paradigms.txt, v_paradigms.txt, tags.txt. There has to be a match between the lemma in the lexicon files, the tag string in the paradigm files and the lemma+tags in the FST. There is a script: scripts/testnouns.sh for testing the generation.

FST

All FSTs are placee at gtlab.uit.no in /opt/smi/crk till we move to gtoahpa.uit.no