Improve locId (#1618) * introduce new model to ontohub * introduce new locid behavior * adding try to prevent undefined method for nil * generate the locids by object creation * introduce method to simulate an attribute * generate Locids correctly for ontologies * correct creation of child_ontology locids * remove old uniqueness validation for ontology locid * use better names vor locidportion * correct behaviour for locid creation * generate loc_id correctly for ontology * use correct class for inherritanc * remove locids on deletion of objects * rename the base model * introduce data-migration to move the locids * remove double naming in the cat ontology * obey hound * correct behaviour of external repository for locid * change factory behaviour for locid * fix missing comma * begin to fix test * change ontology factory to new locid * change factories for new LocId Model * remove unneeded test * Use proper class retrieval in factories. * Always use to_s on the class call in LocIdBaseModel. * Always use .class.to_s instead of .class when used for the database. * Fix external repository. * Be clean and unstub again. * Fix sentence factory. * Quick fix the destruction of locids. * Fix code style. * Generate locids during create process: Base class. * Generate locids during create process: Mapping. * Generate locids during create process: Sentence. * Generate locids during create process: Symbol. * Generate locids during create process: ProofAttempt. * Generate locids during create process: ProverOutput. * Generate locids during create process: Ontology #1. * Generate locids during create process: Ontology #2. * Generate locids during create process: Ontology #3. * Generate locids during create process: Factories. * Fix seeds errors. * Fix ontology spec on error while parsing The ontology needs to be reloaded because sidekiq does so as well. It always fetches a fresh record from the database. The test failed before because `NodeEvaluationHelper#clean_ontology` called destroy on all the ontology's symbols which existed in the object, but were not really persisted because of an aborted transaction. * Remove redundant space. * Move comment to the correct line. * Fix worker_spec In the production code, the locid is never set before saving the ontology. * Add gems pry-stack_explorer, awesome_print for debugging purposes. * Only start hets if there was a hets error and rollback the changes. * Destroy dependent prover output. * Fix ProofAttempt factory. * Rename assorted_object to specific and fix migration. * Fix migration code style. * Reanalyze ontologies with duplicate locids. * Fix code style. * Remove redundancies. * Remove useless association. * Obey Hound. * Directly query in the LocIdBaseModel. * Add locid routing feature. * Cache LocId query results This should speed up the routing in a great extent. For each matching `specified_get`, the LocIdRouterConstraint fetched the element once from the database. This resulted in 8 times the same SQL query for a ProofAttempt. The implemented cache reduces this to a single query. * Clear the elements cache periodically to prevent memory leaks. * Add missing dependent: :destroy. * Add Recreating repository feature. * Combine if statements. * Change Given to When. * Catch ambiguous locid error. * Fix i18n key usage. * Fix code style. * Comment on error suppression. * Add spec for Syntax Error. * Use single-quoted strings. * Fix error message check: language. * Keep parameter assignment DRY. * Rename RecreatingRepository feature to Repository. * Use local_variable_get instead of eval. * Change locid column type to text. * Implement (implicit) down-migration. * Implement (explicit) down migration. * Add migration to create missing locids. * Fix nil error. * Really fix nil error. * Fix migration. * Fix sentence migration accessing association. * Force old Mappings to have a name. * Add missing locids to symbols.

Add missing dependent: :destroy.

Remove unused method prove_options_from_configuration.

Change delegates to has_* ActiveRecord relations.

Remove unused method.

Only preprocess if it has not been preprocessed before.

Revamp the theorem proof status handling.

Use CSAS proof status.

Use ProofExecution instead of CollectiveProofAttempt This allows us to run many proofs in parallel and simplifies the Proof class.

Make a proof attempt retriable.

Add model: TacticScript.

Dasherize locids of ProofAttempt The part after the last // is sort of a name and like Repository#path it should contain dashes instead of being in camel-case.

Add model: ProverOutput.

Use id for routing to a proof attempt again.

Add routes via ProofAttempt locid.

Add locid to ProofAttempt.

Set default proof status of a proof attempt.

Add model: ProofAttemptConfiguration.

Add state to Theorem and ProofAttempt.

Use Prover model in ProofAttempt instead its name.

Use axioms for used_axioms.

Add Numbering to ProofAttempt.

Obey hound.

Wipe theorems for new ontology versions.

Add generated axioms model and create instances When proving, usually a logic translation is applied. Such a translation can introduce sentences as axioms that were not part of the original ontology/specification/theory. Those generated axioms need to be distinguished from the original sentences.

Obey Hound.

Add used_theorems to ProofAttempt Because Theorem is a subclass of Sentence and we have a default scope of Sentence that excludes theorems, we need to have a second relation to get our hands on all the used axioms (sentences and theorems) of a proof.

Add used axioms to proof attempt.

Add ProofStatus model, remove STATUSES Constant. The ProofAttempt and Theorem models now rely on this new model.

Add first draft of Theorem and ProofStatus models.