# Sequence Diagrams
[](https://hackmd.io/eAaJDS6-RDWTPFKD78FH-A)
### ControllerToDesignExtSequence.mmd
sequenceDiagram
autonumber
Controller ->>+ DesignExtractor: extractRelationships(List<HashTable> relationships, List<Node<Statement>> statementList, List<Node<Variable>> variableList)
activate DesignExtractor
Note right of Controller: List<HashTable> relationships includes the following hashtables: <br/> Follows::HashTable<Node<Statement>, Node<Statement>> [Complete]<br/> FollowedBy::HashTable<Node<Statement>, Node<Statement>> [Complete]<br/> Uses::HashTable<Node<Statement>, List<Node<Variable>>> [Partial] <br/> UsedBy::HashTable<Node<Variable>, List<Node<Statement>>> [Partial] <br/> Modifies::HashTable<Node<Statement>, List<Node<Variable>>> [Partial]<br/> ModifiedBy::HashTable<Node<Variable>, List<Node<Statement>>>[Partial] <br/> ParentOf::HashTable<Node<Statement>, List<Node<Statement>>>[Complete] <br/> ChildOf::HashTable<Node<Statement>, List<Node<Statement>>>[Complete]
DesignExtractor ->> DesignExtractor: getTransitiveUses()
DesignExtractor ->> DesignExtractor: getTransitiveUsedBy()
DesignExtractor ->> DesignExtractor: getTransitiveModifies()
DesignExtractor ->> DesignExtractor: getTransitiveModifiedBy()
DesignExtractor ->> DesignExtractor: getParentStarOf()
DesignExtractor ->> DesignExtractor: getChildStarOf()
DesignExtractor -->>+ Controller: List<HashTable> relationships
deactivate DesignExtractor
Note right of Controller: List<HashTable> relationships includes the following hashtables:<br/> Uses::HashTable<Node<Statement>, List<Node<Variable>>> [Complete] <br/> UsedBy::HashTable<Node<Variable>, List<Node<Statement>>> [Complete] <br/> Modifies::HashTable<Node<Statement>, List<Node<Variable>>> [Complete]<br/> ModifiedBy::HashTable<Node<Variable>, List<Node<Statement>>>[Complete] <br/> ParentStarOf::HashTable<Node<Statement>, List<Node<Statement>>>[Complete] <br/> ChildStarOf::HashTable<Node<Statement>, List<Node<Statement>>>[Complete]
Controller ->> PKB: populatePKB(List<HashTable> relationships, List<Node<Statement>> statementList, List<Node<Variable>> variableList)
### ParserToControllerSequence_old.mmd
sequenceDiagram
autonumber
participant AT as AutoTester
participant P as Parser
%% subsystem creates nodes and storing AST while the lines are being read
%% Psystem contains statement stacks..
participant PSub as PSubsystem
participant Tokenizer
participant CGV as ConcreteGrammarValidator
participant C as Controller
%% parser ===> parses the entire program ==> controller
%% parser ===> parses the entire program ==> controller
%% Assumption: This sd only focus on parsing %
AT ->>+ P: parse(sourcefile)
%% Assumption: P will check correctness while building tree %
P ->>+ PSub: createNewTree()
PSub -->>- P: initStatus::IDENTIFIER
%% statement here refers to the source code in between two semicolons
%% If validState == false in reportParseStatus(Failure), the loop terminates
loop hasNextStmt() && validState == True
P ->>+ PSub: sendStatement(stmt)
PSub ->>+ Tokenizer: createTokens(stmt)
Tokenizer ->>+ CGV: checkForValidSyntax(List<TOKEN>)
opt any close brackets
CGV ->>PSub: configStackFlags()
end
CGV -->>- Tokenizer: reportValidationStatus(Boolean)
%% Entity node refers to WHILE/IF/ASSIGN
%% nb: might need to add in a separate node creator here because there will be multiple types of ENTITY_NODE and having an interface here would be better
%% a token looks like: (<symbol>, <tokenTag>) brace_open brace_close
Tokenizer -->>- PSub: ENTITY_NODE
alt ENTITY_NODE INVALID
PSub -->> P: reportParseStatus(INVALID)
else
%% entity is an Abstract/Parent class here,
%% addRelationships modifies the existing DataStructures we have
%% TODO: discuss the logic behind this
PSub ->> PSub: addRelationships(ENTITY_NODE)
PSub -->>- P: reportParseStatus(VALID)
end
end
P -->> AT: programStatus::IDENTIFIER
P -)+ PSub: terminate()
PSub ->>-P: sendDeliverables(AST tree, hashtable, entitytable)
P->>-C: setDataStructures(AST tree, hashtable, entitytable)
%% PSub ->>- C: sendDeliverables(AST tree, hashtable, entitytable)
%% Transitive => Parents*, Children* Follows*, FollowedBy* Use, UsedBy, Modifies, ModifiesBy
%% By right, AST should not be used for resolving queries (anything, we try to add aux Data Structures)
%% This program ends when preprocessing is finish %%
### SourceProcessorToPkbSequence.mmd
sequenceDiagram
autonumber
participant AutoTester
participant Parser
participant Controller
participant DesignExtr
participant PKB
%% Assumption: This sd only focus on parsing %
AutoTester ->> Parser: parse(sourcefile)
% Assumption: Parser will check correctness while building tree %
Parser ->> Parser: createNewTree()
%% todo: extract into separate sequence diagram, representing interactions within the Parser %%
loop hasNextStatement()
Parser -->> Parser: check_create_node()
%%Refer to parser_nodes.mmd for the above statement
Parser -->> Parser: get_relationships_in_line()
loop whileHasRelationship()
Parser -->> Controller: addRelationship(Node n1, Node n2, RSType rs)
end
Parser -->> Controller: addEntity(Node n)
end
Parser ->> Controller: sendAST(AST tree)
%% Transitive => Parents*, Children* Follows*, FollowedBy* Use, UsedBy, Modifies, ModifiesBy
Controller ->> DesignExtr: extractTransitiveRelationships(List<HashTable> relationships, List<Node> entities)
DesignExtr -->> Controller: return
Controller ->> PKB: storeAST_HashTable_Entity(AST tree, List<HashTable> ht, List<Node> entities)
%% By right, AST should not be used for resolving queries (anything, we try to add aux Data Structures)
Parser -->> AutoTester: parseStatus::int
%% This program ends when preprocessing is finish %%
### SourceProcessorToPkbSequence_detailed.mmd
sequenceDiagram
autonumber
participant AT as AutoTester
participant SP as SourceProcessor
participant P as Parser
%% subsystem creates nodes and storing AST while the lines are being read
%% Psystem contains statement stacks..
participant PSub as PSubsystem
participant Tkn as Tokenizer
participant CGV as ConcreteGrammarValidator
%% EntityFactory modifies necessary DSes (creates nodes in AST, add entries into relationship, entity tables ...)
%% To prevent the passing of multiple different objects, there's going to be a bi-directional relationship b/w EF and PSub where PSub will read the values of DSes and Helper Stacks kept inside PSub, and EF will be creating and modifying said objects within PSub. This might need an intermediary Relationship Class or smth to avoid circular dependency
participant EF as EntityFactory
participant DE as DesignExtractor
participant PKB as ProgramKnowledgeBase
AT->>+SP: processSourceFile(sourceFile)
SP ->>+ P: parse(sourcefile)
%% initialises the necessary data structures inside PSub, these DSes (AST, RelationshipTables, EntityTables) will be accumulated
%% Assumption: P will check correctness while building DSes %
P ->>+ PSub: initDataStructures()
PSub -->>- P: initStatus::IDENTIFIER
%% statement here refers to the source code in between two semicolons
%% If validState == false in reportParseStatus(Failure), the loop terminates
loop hasNextStmt && validState
P ->>+ PSub: processStatement(stmt)
PSub ->>+ Tkn: createTokens(stmt)
%% a token looks like: (<symbol>, <tokenTag>) brace_open brace_close
Tkn -->>- PSub: tokens::List<TOKEN>
%% We ask the validator to validate the syntax for a single statement, if there's an error at any token, it should fail early and return us the validationStatus
%% SideEffect:Since the validator looks at tokens one by one, we let the validator modify "helper stacks" that are in the PSub
%% "Helper Stacks" are cumulative stacks or any other DSes kept in the PSubsystem. e.g.
%% "Helper Stacks" are there to help keep track of tokens and determine how the DSes are being modified. e.g. if CGV current sees an open brace, it knows that it's a new container, hence the PSub's helper stack should be notified such that the AST will branch out, the relationship tables will create new entries...
PSub ->>+ CGV: validateConcreteSyntax(List<TOKEN>)
CGV -->>-PSub: validationStatus::IDENTIFIER
%% create Entities: Entity here refers to the Abstraction Entity as described by the wiki/lecture 1 slides
PSub ->>+EF:createEntities(List<Tokens>)
%% This method has responsibilities such as adding necessary Entity Nodes to the AST and updating the necessary tables to reflect relationships (this is where we add info about relationships b/w lines, vars...)
%% Entity node refers to WHILE/IF/ASSIGN... ref to
EF-->>-PSub:completionStatus::IDENTIFIER
PSub -->>-P: statementProcessStatus::IDENTIFIER
%% todo: handle the error handling opt code here
end
%% at this point, all the statements have been processed, the parser is ready for deliverables to be sent to SP
P ->>+PSub: sendDeliverables()
PSub -->>-P: deliverables::DELIVERABLES
P -->>-SP: parseStatus::IDENTIFIER
SP ->>+P: sendDeliverables()
P -->>-SP: deliverables::DELIVERABLES
%% todo: add in DE activities
SP ->>+DE:extractDesigns()
%% DE and SP will comms with each other bidirectionally
%% todo: these sequences will be fleshed out in a separate sequence diagram
DE -->>-SP:extractionStatus::IDENTIFIER
alt valid sourceProcessStatus
SP ->>+PKB: populateDataStructures()
PKB -->>-SP:pkbSuccessStatus::IDENTIFIER
else
SP ->>SP : handleSourceCodeError()
end
SP -->>- AT: sourceProcessStatus::IDENTIFIER
%% Transitive => Parents*, Children* Follows*, FollowedBy* Use, UsedBy, Modifies, ModifiesBy
%% By right, AST should not be used for resolving queries (anything, we try to add aux Data Structures)
%% This sequence of steps ends when the source processing is done and deliverables sent to PKB %%
### queries_overview.mmd
sequenceDiagram
autonumber
participant QueryExtractor
participant QueryParser
participant QueryTokenizer
participant QuerySemanticValidator
participant QueryGrouper
%% This sd is for per query (PQL statement) basis.
%% TODO: include initialiser
QueryExtractor->>QueryTokenizer: init()
QueryTokenizer-->>QueryExtractor: QueryTokenizer tokenizer
QueryExtractor->>QueryParser: init(vector<Clause*> clauses, list<Synonym*> synonyms, <br>vector<Synonym*> target_synonyms, tokenizer)
QueryParser-->>QueryExtractor: QueryParser parser
QueryExtractor->>QueryParser: parser.Parse()
Note right of QueryParser: Calls multiple handlers according to <br> grammar rules (recursive descent)
loop is expecting tokens based on grammar rules
QueryParser->>QueryTokenizer: GetNextToken()
QueryTokenizer-->>QueryParser: Token t
Note right of QueryParser: First layer of validations
opt is relRef
Note right of QueryParser: Additional layer of validations
QueryParser->>QuerySemanticValidator: Is_Semantically_Valid_RelRef()
QuerySemanticValidator-->>QueryParser: Boolean b
end
opt is attrRef
Note right of QueryParser: Additional layer of validations
QueryParser->>QuerySemanticValidator: Is_Semantically_Valid_AttrRef()
QuerySemanticValidator-->>QueryParser: Boolean b
end
end
QueryParser-->>QueryExtractor:
QueryExtractor->>QueryExtractor: PopulateSynAdjacencyList(<br>Map<string, vector<int>* adj_list, <br>vector<Clause*> clauses)
Note right of QueryExtractor: Populates data structure for QueryGrouper
QueryExtractor->>QueryGrouper: GroupClauses(clauses, groups, target_synonyms, adj_list)
QueryGrouper-->>QueryExtractor:
### queries.mmd (deprecated)
sequenceDiagram
autonumber
participant AutoTester
participant PqlPreprocessor
participant QueryEvaluator
participant PKB
participant QueryResultProjector
participant UI
%% This sd is for per query (PQL statement) basis.
AutoTester ->> PqlPreprocessor: evaluate(q)
PqlPreprocessor->>PqlPreprocessor: tokenizePql(q)
PqlPreprocessor->>PqlPreprocessor: parsePql(q)
PqlPreprocessor->>PqlPreprocessor: verifyPql(q)
PqlPreprocessor->>PqlPreprocessor: orderClauses()
PqlPreprocessor->>QueryEvaluator: sendQueryStruct(q)
%% Refer to query_evaluation_pkb.mmd
note right of PKB: info refers to relevant PSub-structures <br/> (maybe a linked list...) that <br/> the QueryEvaluatorcan use <br/> to resolve the query.
QueryEvaluator->>PKB: retrieveClauseInfo(clause)
%% MergeSets() represents union, intersect, join between the different clauses results.
QueryEvaluator ->>QueryEvaluator:mergeSets()
QueryEvaluator ->>QueryResultProjector: generateResultFormatter(RESULT r)
QueryResultProjector ->> UI: finalResult(FORMATTED_RESULT fr)
### query_evaluation_pkb.mmd
sequenceDiagram
participant QueryPreProcessor
participant QueryEvaluator
participant PKB
participant QueryResultProjector
% synList - all syn declared
% e.g assign a; variable v;
% Select v such that Parent (a, 1) pattern a ("x", "1 + 2")
QueryPreProcessor->>+QueryEvaluator: QueryStruct(SYNONYM[] synList, SYNONYM target, SUCH_THAT[] stList, PATTERN[] patternList)
QueryEvaluator->>+PKB: getProcTable()
PKB-->>-QueryEvaluator: PROC[]
% querying smts or var is not put into an alt else to illustrate extension for procedure and call queries later on
loop Every valid synonym declaration
opt querying statements
% Note that Stmt can be ‘stmt’ | ‘read’ | ‘print’ | ‘while’ | ‘if’ | ‘assign’
QueryEvaluator->>+PKB: getStmtTable(STMT_TYPE stmtType)
PKB-->>-QueryEvaluator: STMT[] listOfStmt, ERR_CODE err
end
opt querying variables
% Note that Var can be ‘variable’ | ‘constant’
QueryEvaluator->>+PKB: getVarTable(VAR_TYPE varType)
PKB-->>-QueryEvaluator: STMT[] listOfVar, ERR_CODE err
end
end
% For each such that or pattern
loop Every such that clause
opt Select s such that Follows (stmt1, s)
QueryEvaluator->>+PKB: getFollows (STMT stmt1)
PKB-->>-QueryEvaluator: STMT clauseResult
end
opt Select s such that Follows* (stmt1, s)
QueryEvaluator->>+PKB: getFollowsStar (STMT stmt1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
opt Select s such that Follows (s, stmt1)
QueryEvaluator->>+PKB: getPrevious (STMT stmt1)
PKB-->>-QueryEvaluator: STMT clauseResult
end
opt Select s such that Follows* (s, stmt1)
QueryEvaluator->>+PKB: getPreviousStar (STMT stmt1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
opt Select s such that Parent (s, stmt1)
QueryEvaluator->>+PKB: getParent (STMT stmt1)
PKB-->>-QueryEvaluator: STMT clauseResult
end
opt Select s such that Parent* (s, stmt1)
QueryEvaluator->>+PKB: getParentStar (STMT stmt1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
opt Select s such that Parent (stmt1, s)
QueryEvaluator->>+PKB: getChild (STMT stmt1)
PKB-->>-QueryEvaluator: STMT clauseResult
end
opt Select s such that Parent* (stmt1, s)
QueryEvaluator->>+PKB: getChildStar (STMT stmt1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
opt Select v such that Uses (stmt1, v)
QueryEvaluator->>+PKB: getVarUsed (STMT stmt1)
PKB-->>-QueryEvaluator: VAR clauseResult
end
opt Select s such that Uses (s, var1)
QueryEvaluator->>+PKB: getStmtUsing (VAR var1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
opt Select v such that Modifies (stmt1, v)
QueryEvaluator->>+PKB: getVarModified (STMT stmt1)
PKB-->>-QueryEvaluator: VAR clauseResult
end
opt Select s such that Modifies (s, var1)
QueryEvaluator->>+PKB: getStmtModifying (VAR var1)
PKB-->>-QueryEvaluator: STMT[] clauseResult
end
end
opt has >= 1 pattern-cl && notCached
QueryEvaluator ->> PKB: getAllAssignStmts()
PKB -->> QueryEvaluator: AST_ASSIGN[] assignStmts
QueryEvaluator ->> QueryEvaluator: cacheAssignStmts
end
loop Each pattern clause
QueryEvaluator->>QueryEvaluator: STMT[] comparePattern(PATTERN p)
note right of QueryEvaluator: Case lhs: <br/>if "var_name" (exact match): filter assignStmts<br/>if synonym: variable filter assignStmts<br/>if _ (any match): do nothing
note right of QueryEvaluator: Case rhs: <br/> if "..." (exact match): filter assignStmts<br/>if _"..."_ (partial match): check AST_ASSIGNS for subtree match first<br/>if _ (any match): do nothing
end
% Consider extracting the work of combining clauses to another
% component.
% Compute combination of clause results
opt querying statements
QueryEvaluator->>+QueryEvaluator: STMT[] ComputeCombinedResult(STMT[][])
end
opt querying variables
QueryEvaluator->>+QueryEvaluator: VAR[] ComputeCombinedResult(VAR[][])
end
QueryEvaluator->>QueryResultProjector: STMT[]/VAR[] RESULT
Sign in with Wallet
Connect another wallet