diff options
Diffstat (limited to 'subex/subexast.go')
-rw-r--r-- | subex/subexast.go | 468 |
1 files changed, 315 insertions, 153 deletions
diff --git a/subex/subexast.go b/subex/subexast.go index e02091d..655a783 100644 --- a/subex/subexast.go +++ b/subex/subexast.go @@ -7,54 +7,72 @@ import ( // A node in the AST of a subex type SubexAST interface { - compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState + compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState } // Process the first subex, then the second, splitting the input text in two type SubexASTConcat struct { First, Second SubexAST } -func (ast SubexASTConcat) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - return ast.First.compileWith(ast.Second.compileWith(next, slotMap, runic), slotMap, runic) +func (ast SubexASTConcat) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + return ast.First.compileWith( + ast.Second.compileWith(next, slotMap, inType, outType), + slotMap, + inType, + outType, + ) } func (ast SubexASTConcat) String() string { return fmt.Sprintf("(%v)(%v)", ast.First, ast.Second) } // Processing a subex and storing the output in a slot instead of outputting it -type SubexASTStore struct { +type SubexASTStoreValues struct { Match SubexAST Slot rune } -func (ast SubexASTStore) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTStoreValues) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { id := slotMap.getId(ast.Slot) newNext := ast.Match.compileWith(&SubexStoreEndState { slot: id, next: next, - }, slotMap, runic) + }, slotMap, inType, ValueType) - if !runic { - return &SubexCaptureBeginState { - next: newNext, - } - } else { - return &SubexCaptureRunesBeginState { - next: newNext, - } + return &SubexCaptureBeginState { + next: newNext, } } -func (ast SubexASTStore) String() string { +func (ast SubexASTStoreValues) String() string { return fmt.Sprintf("$%c(%v)", ast.Slot, ast.Match) } +type SubexASTStoreRunes struct { + Match SubexAST + Slot rune +} +func (ast SubexASTStoreRunes) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + id := slotMap.getRuneId(ast.Slot) + newNext := ast.Match.compileWith(&SubexStoreRunesEndState { + slot: id, + next: next, + }, slotMap, inType, RuneType) + + return &SubexCaptureRunesBeginState { + next: newNext, + } +} +func (ast SubexASTStoreRunes) String() string { + return fmt.Sprintf("(%v)$%c", ast.Match, ast.Slot) +} + // Try to run the first subex, if it fails then backtrack and use the second type SubexASTOr struct { First, Second SubexAST } -func (ast SubexASTOr) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTOr) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { return &SubexGroupState { - ast.First.compileWith(next, slotMap, runic), - ast.Second.compileWith(next, slotMap, runic), + ast.First.compileWith(next, slotMap, inType, outType), + ast.Second.compileWith(next, slotMap, inType, outType), } } func (ast SubexASTOr) String() string { @@ -84,19 +102,24 @@ func (cr ConvexRange) decrement() ConvexRange { return ConvexRange{cr.Start - 1, cr.End - 1} } } -func (cr ConvexRange) compile(content SubexAST, next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (cr ConvexRange) compile(content SubexAST, next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { min, _ := cr.minmax() if min != 0 { - return content.compileWith(cr.decrement().compile(content, next, slotMap, runic), slotMap, runic) + return content.compileWith( + cr.decrement().compile(content, next, slotMap, inType, outType), + slotMap, + inType, + outType, + ) } if cr.Start == -1 { state := &SubexGroupState {nil, next} - state.first = content.compileWith(state, slotMap, runic) + state.first = content.compileWith(state, slotMap, inType, outType) return state } if cr.End == -1 { state := &SubexGroupState {next, nil} - state.second = content.compileWith(state, slotMap, runic) + state.second = content.compileWith(state, slotMap, inType, outType) return state } @@ -104,7 +127,7 @@ func (cr ConvexRange) compile(content SubexAST, next SubexState, slotMap *SlotMa state := next; for i := 0; i < cr.Start; i += 1 { state = &SubexGroupState { - content.compileWith(state, slotMap, runic), + content.compileWith(state, slotMap, inType, outType), next, } } @@ -114,7 +137,7 @@ func (cr ConvexRange) compile(content SubexAST, next SubexState, slotMap *SlotMa for i := 0; i < cr.End; i += 1 { state = &SubexGroupState { next, - content.compileWith(state, slotMap, runic), + content.compileWith(state, slotMap, inType, outType), } } return state @@ -127,10 +150,10 @@ type SubexASTRepeat struct { Content SubexAST Acceptable []ConvexRange } -func (ast SubexASTRepeat) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTRepeat) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { var state SubexState = &SubexDeadState{} for _, convex := range ast.Acceptable { - state = &SubexGroupState {state, convex.compile(ast.Content, next, slotMap, runic)} + state = &SubexGroupState {state, convex.compile(ast.Content, next, slotMap, inType, outType)} } return state } @@ -142,7 +165,10 @@ func (ast SubexASTRepeat) String() string { type SubexASTCopyScalar struct { Scalar walk.Scalar } -func (ast SubexASTCopyScalar) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyScalar) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTCopyScalar") + } return &SubexCopyState{ filter: selectScalarFilter {ast.Scalar}, next: next, @@ -153,26 +179,41 @@ func (ast SubexASTCopyScalar) String() string { } type SubexASTCopyAnyRune struct {} -func (ast SubexASTCopyAnyRune) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyAnyRune) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != RuneType || outType != RuneType { + panic("Invalid types for SubexASTCopyAnyRune") + } return &SubexCopyRuneState { next: next, filter: anyRuneFilter{}, } } +func (ast SubexASTCopyAnyRune) String() string { + return ".RUNE" +} type SubexASTCopyRune struct { rune rune } -func (ast SubexASTCopyRune) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyRune) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != RuneType || outType != RuneType { + panic("Invalid types for SubexASTCopyRune") + } return &SubexCopyRuneState { next: next, filter: selectRuneFilter {ast.rune}, } } +func (ast SubexASTCopyRune) String() string { + return string(ast.rune) +} // Read in a single atom that must be a boolean and output it unchanged type SubexASTCopyBool struct {} -func (ast SubexASTCopyBool) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyBool) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTCopyBool") + } return &SubexCopyState { next: next, filter: anyBoolFilter{}, @@ -184,7 +225,10 @@ func (ast SubexASTCopyBool) String() string { // Read in a single atom that must be a number and output it unchanged type SubexASTCopyNumber struct {} -func (ast SubexASTCopyNumber) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyNumber) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTCopyNumber") + } return &SubexCopyState { next: next, filter: anyNumberFilter{}, @@ -194,34 +238,24 @@ func (ast SubexASTCopyNumber) String() string { return "%" } -// Read in a full string value and copy it out unchanged -// # is equivalent to "_{-0}" -// TODO -// type SubexASTCopyString struct {} -// func (ast SubexASTCopyString) compileWith(next SubexState, slotMap *SlotMap) SubexState { -// stringAtomState := &SubexCopyStringAtomState { -// next: nil, -// } -// stringContentState := &SubexGroupState { -// &SubexCopyScalarState { -// scalar: walk.NewAtomStringTerminal(), -// next: next, -// }, -// stringAtomState, -// } -// stringAtomState.next = stringContentState -// return &SubexCopyScalarState { -// scalar: walk.NewAtomStringTerminal(), -// next: stringContentState, -// } -// } -// func (ast SubexASTCopyString) String() string { -// return "#" -// } +// Read in a null, bool, number, string or empty array or map and output it unchanged +type SubexASTCopyAnySimpleValue struct {} +func (ast SubexASTCopyAnySimpleValue) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTCopyAnySimpleValue") + } + return &SubexCopyState { + next: next, + filter: simpleValueFilter{}, + } +} // Read in any single Atom and output it unchanged type SubexASTCopyAnyValue struct {} -func (ast SubexASTCopyAnyValue) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTCopyAnyValue) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTCopyAnyValue") + } return &SubexCopyState { next: next, filter: anyValueFilter{}, @@ -231,6 +265,7 @@ func (ast SubexASTCopyAnyValue) String() string { return "." } +/* type OutputContentAST interface { compile(slotMap *SlotMap) OutputContent } @@ -273,25 +308,58 @@ func (ast SubexASTOutput) compileWith(next SubexState, slotMap *SlotMap, runic b func (ast SubexASTOutput) String() string { return "=...=" } +*/ -// Read in a repeated subex separated by a delimiter. Greedy -type SubexASTJoin struct { - Content, Delimiter SubexAST +type SubexASTOutputValueLiteral struct { + literal walk.Scalar } -func (ast SubexASTJoin) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - afterContentState := &SubexGroupState { - nil, - next, +func (ast SubexASTOutputValueLiteral) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if outType != ValueType { + panic("Invalid outType for SubexASTOutputValueLiteral") } - manyContentsState := ast.Content.compileWith(afterContentState, slotMap, runic) - afterContentState.first = ast.Delimiter.compileWith(manyContentsState, slotMap, runic) - return &SubexGroupState { - manyContentsState, - next, + return &SubexOutputValueLiteralState { + literal: ast.literal, + next: next, + } +} + +type SubexASTOutputValueLoad struct { + slot rune +} +func (ast SubexASTOutputValueLoad) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if outType != ValueType { + panic("Invalid outType for SubexASTOutputValueLoad") + } + return &SubexOutputValueLoadState { + slot: slotMap.getId(ast.slot), + next: next, + } +} + +type SubexASTOutputRuneLiteral struct { + literal rune +} +func (ast SubexASTOutputRuneLiteral) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if outType != RuneType { + panic("Invalid outType for SubexASTOutputRuneLiteral") + } + return &SubexOutputRuneLiteralState { + literal: ast.literal, + next: next, } } -func (ast SubexASTJoin) String() string { - return fmt.Sprintf("(%v);(%v)", ast.Content, ast.Delimiter) + +type SubexASTOutputRuneLoad struct { + slot rune +} +func (ast SubexASTOutputRuneLoad) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if outType != RuneType { + panic("Invalid outType for SubexASTOutputRuneLoad") + } + return &SubexOutputRuneLoadState { + slot: slotMap.getRuneId(ast.slot), + next: next, + } } // Run each input Atom through a map to produce an output Atom @@ -314,12 +382,15 @@ func (ast SubexASTJoin) String() string { type SubexASTSum struct { Content SubexAST } -func (ast SubexASTSum) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTSum) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTSum") + } return &SubexCaptureBeginState { next: ast.Content.compileWith(&SubexArithmeticEndState { next: next, calculate: sumValues, - }, slotMap, runic), + }, slotMap, inType, outType), } } func (ast SubexASTSum) String() string { @@ -330,12 +401,15 @@ func (ast SubexASTSum) String() string { type SubexASTProduct struct { Content SubexAST } -func (ast SubexASTProduct) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTProduct) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTProduct") + } return &SubexCaptureBeginState { next: ast.Content.compileWith(&SubexArithmeticEndState { next: next, calculate: multiplyValues, - }, slotMap, runic), + }, slotMap, inType, outType), } } func (ast SubexASTProduct) String() string { @@ -347,12 +421,15 @@ func (ast SubexASTProduct) String() string { type SubexASTNegate struct { Content SubexAST } -func (ast SubexASTNegate) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTNegate) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTNegate") + } return &SubexCaptureBeginState { next: ast.Content.compileWith(&SubexArithmeticEndState { next: next, calculate: negateValues, - }, slotMap, runic), + }, slotMap, inType, outType), } } func (ast SubexASTNegate) String() string { @@ -360,58 +437,29 @@ func (ast SubexASTNegate) String() string { } // Runs the content Subex and collects the output -// If it is a list of atoms castable to numbers, it takes the reciprocal of them all and outputs them -// Else it rejects -type SubexASTReciprocal struct { - Content SubexAST -} -func (ast SubexASTReciprocal) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - return &SubexCaptureBeginState { - next: ast.Content.compileWith(&SubexArithmeticEndState { - next: next, - calculate: reciprocalValues, - }, slotMap, runic), - } -} -func (ast SubexASTReciprocal) String() string { - return fmt.Sprintf("(%v)/", ast.Content) -} - -// Runs the content Subex and collects the output // Maps over the values in the output, casting each to a boolean, notting each and then outputs them // Rejects if it cannot cast to boolean type SubexASTNot struct { Content SubexAST } -func (ast SubexASTNot) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTNot) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + if inType != ValueType || outType != ValueType { + panic("Invalid types for SubexASTNot") + } return &SubexCaptureBeginState { next: ast.Content.compileWith(&SubexArithmeticEndState { next: next, calculate: notValues, - }, slotMap, runic), + }, slotMap, ValueType, ValueType), } } func (ast SubexASTNot) String() string { return fmt.Sprintf("(%v)!", ast.Content) } -// Runs the content Subex and collects the output -// Replaces it with true if all output values are equal and false otherwise -type SubexASTEqual struct { - Content SubexAST -} -func (ast SubexASTEqual) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - return &SubexCaptureBeginState { - next: ast.Content.compileWith(&SubexArithmeticEndState { - next: next, - calculate: equalValues, - }, slotMap, runic), - } -} - // Does nothing type SubexASTEmpty struct {} -func (ast SubexASTEmpty) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { +func (ast SubexASTEmpty) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { return next } func (ast SubexASTEmpty) String() string { @@ -421,10 +469,11 @@ func (ast SubexASTEmpty) String() string { // Discards the output from the content subex type SubexASTDiscard struct { Content SubexAST + InnerOutType Type } -func (ast SubexASTDiscard) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - newNext := ast.Content.compileWith(&SubexDiscardState {next}, slotMap, runic) - if !runic { +func (ast SubexASTDiscard) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + newNext := ast.Content.compileWith(&SubexDiscardState {next}, slotMap, inType, ast.InnerOutType) + if inType == ValueType { return &SubexCaptureBeginState { next: newNext, } @@ -438,57 +487,170 @@ func (ast SubexASTDiscard) String() string { return fmt.Sprintf("(%v)$_", ast.Content) } -// Go into an array, pass the content each of the values in the array to eat and then leave the array -type SubexASTEnterArray struct { +type SubexASTDestructure struct { + Destructure Structure + Structure Structure Content SubexAST } -func (ast SubexASTEnterArray) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - return &SubexCaptureBeginState { - next: &SubexIncrementNestState { +func (ast SubexASTDestructure) compileWith(next SubexState, slotMap *SlotMap, inType Type, outType Type) SubexState { + var innerOutType Type + var construct SubexState + switch ast.Structure { + case NoneStructure: + innerOutType = outType + construct = next + case StringStructure: + innerOutType = RuneType + construct = &SubexConstructStringState { + next: next, + } + case ArrayStructure: + innerOutType = ValueType + construct = &SubexConstructArrayState { + next: next, + } + case ArrayValuesStructure: + innerOutType = ValueType + construct = &SubexConstructArrayValuesState { + next: next, + } + case MapStructure: + innerOutType = ValueType + construct = &SubexConstructMapState { + next: next, + } + default: + panic("Invalid ast structure") + } + + var innerInType Type + var destructFooter SubexState + switch ast.Destructure { + case NoneStructure: + innerInType = inType + destructFooter = construct + case StringStructure: + innerInType = RuneType + destructFooter = &SubexDiscardTerminalState { + terminal: walk.StringEnd, + next: &SubexDecrementNestState { + next: construct, + }, + } + case ArrayStructure: + innerInType = ValueType + destructFooter = &SubexDiscardTerminalState { + terminal: walk.ArrayEnd, + next: &SubexDecrementNestState { + next: construct, + }, + } + case ArrayValuesStructure: + innerInType = ValueType + destructFooter = &SubexDiscardTerminalState { + terminal: walk.ArrayEnd, + next: &SubexDecrementNestState { + next: construct, + }, + } + case MapStructure: + innerInType = ValueType + destructFooter = &SubexDiscardTerminalState { + terminal: walk.MapEnd, + next: &SubexDecrementNestState { + next: construct, + }, + } + default: + panic("Invalid ast destructure") + } + + inner := ast.Content.compileWith( + destructFooter, + slotMap, + innerInType, + innerOutType, + ) + + var beginConstruct SubexState + switch ast.Structure { + case NoneStructure: + beginConstruct = inner + case StringStructure: + beginConstruct = &SubexCaptureRunesBeginState { + next: inner, + } + case ArrayStructure: + beginConstruct = &SubexCaptureBeginState { + next: inner, + } + case ArrayValuesStructure: + beginConstruct = &SubexCaptureBeginState { + next: inner, + } + case MapStructure: + beginConstruct = &SubexCaptureBeginState { + next: inner, + } + default: + panic("Invalid ast structure") + } + + switch ast.Destructure { + case NoneStructure: + return beginConstruct + case StringStructure: + return &SubexCaptureBeginState { + next: &SubexCopyState { + filter: anyStringFilter{}, + next: &SubexDiscardState { + next: &SubexIncrementNestState { + keys: true, + next: beginConstruct, + }, + }, + }, + } + case ArrayStructure: + return &SubexCaptureBeginState { next: &SubexCopyState { filter: anyArrayFilter{}, next: &SubexDiscardState { - next: &SubexCaptureBeginState { - next: ast.Content.compileWith( - &SubexDiscardTerminalState { - terminal: walk.ArrayEndTerminal{}, - next: &SubexDecrementNestState { - next: &SubexConstructArrayState {next: next}, - }, - }, - slotMap, - runic, - ), + next: &SubexIncrementNestState { + keys: true, + next: beginConstruct, }, }, }, - }, - } -} - -type SubexASTEnterString struct { - Content SubexAST -} -func (ast SubexASTEnterString) compileWith(next SubexState, slotMap *SlotMap, runic bool) SubexState { - return &SubexCaptureBeginState { - next: &SubexIncrementNestState { + } + case ArrayValuesStructure: + return &SubexCaptureBeginState { next: &SubexCopyState { - filter: anyStringFilter{}, + filter: anyArrayFilter{}, + next: &SubexDiscardState { + next: &SubexIncrementNestState { + keys: false, + next: beginConstruct, + }, + }, + }, + } + case MapStructure: + return &SubexCaptureBeginState { + next: &SubexCopyState { + filter: anyMapFilter{}, next: &SubexDiscardState { - next: &SubexCaptureRunesBeginState { - next: ast.Content.compileWith( - &SubexDecrementNestState { - next: &SubexDiscardTerminalState { - terminal: walk.StringEndTerminal{}, - next: &SubexConstructStringState {next: next}, - }, - }, - slotMap, - true, - ), + next: &SubexIncrementNestState { + keys: true, + next: beginConstruct, }, }, }, - }, + } + default: + panic("Invalid destructure in ast") } } +func (ast SubexASTDestructure) String() string { + return fmt.Sprintf("%v(%v)%v", ast.Destructure, ast.Content, ast.Structure) +} |