2022-03-01

[turbomam]

• [ ] want modeling and tools for
  • [ ] different designation for genes, proteins and enzymes including the organism they come from. See Gene to Protein encoded
  • [ ] types of modifications like fluorescent/epitope
• [ ] heterogeneous enums like modification_type_enum, mod_category_enum and genomic_footprint_enum for descriptor... model as classes in addition or instead of enumerations?
• [ ] what terms are required?
• [ ] what is the difference between a modification, part and element
• [ ] ranges and domains don't appear in owl
• [ ] permissible values appear as classes in owl/Protege, with the enum as a union class
• [ ] slot uris break something?
  • class_uris seem OK
• [ ] examples via schema sheets
• [ ] aliases for (organism) PVs
• [ ] log field names that require quoting
• [ ] can mixins have is_a parents?
• [ ] patterns on inherited IDs: ValueError: Class "Part" - multiple keys/identifiers not allowed (id, Part_id)
• [ ] remove definitions of terms that are defined in the metamodel like alias, comment...
• [ ] mangled special characters in DH: Part:14's genotype_phenotype = MATalpha his3Δ1 leu2Δ0 met15Δ0 ura3Δ0
  • [ ] check the YAML too
• [ ] gen-markdown/mkdocs
  • [ ] search not working
  • [ ] index page not self evident: http://127.0.0.1:8000/synbio_schema_sheet/
• [x] boolean attributes like circular
• [x] pipe separated aliases
• [x] reinstate DH interface, with section and slot order annotations

Ignored tables despite links to parts

• empty
  • parts_comments
  • arabidopsis_seeds
  • proteins
• ids only
  • strains
• populated
  • biological_samples

[turbomam]

Notes from SQL queries

• which are REQUIRED?
• keeping references and permissible values semantic
• taxon ids should be as specific AS POSSIBLE
  • find a way to represent non-phylogenetic groups
• when enums and when subclasses? think about the semantics and about data entry user experience
• include definitions
• define classes with hierarchy and inter-relations
• minimal number of slots/fields/columns with reuse and class-slot-usage
• controlled vocabularies should be topically consistent
• tooling: data entry, conversion, repair, validation
  • validation can include internal consistency checks (NCBI_taxon vs binomial name)

Modifications

• scalar attributes

  • "element" -> element_long | element_name -> element_short | "size" -> size_bp | "position" -> modification_position | aa_change | subcategory_size |

• enums

  • "descriptor" -> genomic_footprint | modification_type | category -> modification_category |

• just one prefixed ID... modification:IF...?

  • element_id_number | element_id |
  • where to assert parrerns (or prefix patterns?) for ids? Also DH sections and column orders?
  • Can't do it in slot usage, whcih complains about multiple primary keys

• parts_ptr_id | model as has_modifications relationship FROM Parts

  • how to check prefixes/patterns of mentioned things by ID?

• model with one link to an Organism. How to make the identifiers semantic?

  • element_species | taxon_id

Parts

• ID

  • id |

• enums

  • "type" -> part_type | funding_source | status -> part_status | bio_safety_level | intellectual_property

• "name" | alias -> aliases | genotype_phenotype | keywords | summary | "references" -> lit_references | notes |

• links to People with pi and creator predicates

  • principal_investigator | principal_investigator_email | creator_id |

• links to Organisms with organism_basis predicates

  • host_species_id

[turbomam]

• [x] Add reasoning test over owl ttt
  • [ ] or at least text search for "error"

[turbomam]

id and name columns/section should come first in DH

DH should look at schema.slots if there's no section or column order annotation on an induced slot.

[turbomam]

restore these as required on some parts

• bio_safety_level
• creator
• part_type
• pi
• status

[turbomam]

unique without setting identifier?

[turbomam]

test documentation and DH with selenium ?!