package uri import ( "fmt" "strings" ) // === WILDCARDS === // // The star '*' -> matches 0 or 1 slash-bounded string // The multi star '**' -> matches 0 or more slash-separated strings // The dot '.' -> matches 1 slash-bounded string // The multi dot '..' -> matches 1 or more slash-separated strings // // === SCHEME POLICY === // // - The last '/' is optional // - Any '**' at the very end will match anything that starts with the given prefix // // === LIMITATIONS == // // - A scheme must begin with '/' // - A scheme cannot contain something else than a STRING or WILDCARD between 2 '/' separators // - A scheme STRING cannot contain the symbols '/' as a character // - A scheme STRING containing '*' or '.' characters will be treating as STRING only // - A maximum of 16 slash-separated matchers (STRING or WILDCARD) are allowed const maxMatch = 16 // Represents an URI matcher type matcher struct { pat string // pattern to match (empty if wildcard) req bool // whether it is required mul bool // whether multiple matches are allowed buf []string // matched content (when matching) } // Scheme represents an URI scheme type Scheme []*matcher // FromString builds an URI scheme from a pattern string func FromString(s string) (*Scheme, error) { // 1. Manage '/' at the start if len(s) < 1 || s[0] != '/' { return nil, fmt.Errorf("URI must begin with '/'") } // 2. Split by '/' parts := strings.Split(s, "/") // 3. Max exceeded if len(parts)-2 > maxMatch { for i, p := range parts { fmt.Printf("%d: '%s'\n", i, p) } return nil, fmt.Errorf("URI must not exceed %d slash-separated components, got %d", maxMatch, len(parts)) } // 4. Build for each part sch, err := buildScheme(parts) if err != nil { return nil, err } // 5. Optimise structure opti, err := sch.optimise() if err != nil { return nil, err } return &opti, nil } // Match returns if the given URI is matched by the scheme func (s Scheme) Match(str string) bool { // 1. Nothing -> match all if len(s) == 0 { return true } // 2. Check for string match clearURI, match := s.matchString(str) if !match { return false } // 3. Check for non-string match (wildcards) match = s.matchWildcards(clearURI) if !match { return false } return true } // GetMatch returns the indexed match (excluding string matchers) func (s Scheme) GetMatch(n uint8) ([]string, error) { // 1. Index out of range if n > uint8(len(s)) { return nil, fmt.Errorf("Index out of range") } // 2. Iterate to find index (exclude strings) ni := -1 for _, m := range s { // ignore strings if len(m.pat) > 0 { continue } // increment match counter : ni ni++ // if expected index -> return matches if uint8(ni) == n { return m.buf, nil } } // 3. If nothing found -> return empty set return nil, fmt.Errorf("Index out of range (max: %d)", ni) } // GetAllMatch returns all the indexed match (excluding string matchers) func (s Scheme) GetAllMatch() [][]string { match := make([][]string, 0, len(s)) for _, m := range s { // ignore strings if len(m.pat) > 0 { continue } match = append(match, m.buf) } return match } // buildScheme builds a 'basic' scheme // from a pattern string func buildScheme(ss []string) (Scheme, error) { // 1. Build scheme sch := make(Scheme, 0, maxMatch) for _, s := range ss { // 2. ignore empty if len(s) == 0 { continue } m := &matcher{} switch s { // 3. Card: 0, N case "**": m.req = false m.mul = true sch = append(sch, m) // 4. Card: 1, N case "..": m.req = true m.mul = true sch = append(sch, m) // 5. Card: 0, 1 case "*": m.req = false m.mul = false sch = append(sch, m) // 6. Card: 1 case ".": m.req = true m.mul = false sch = append(sch, m) // 7. Card: 1, literal string default: m.req = true m.mul = false m.pat = fmt.Sprintf("/%s", s) sch = append(sch, m) } } return sch, nil } // optimise optimised the scheme for further parsing func (s Scheme) optimise() (Scheme, error) { // 1. Nothing to do if only 1 element if len(s) <= 1 { return s, nil } // 2. Init reshifted scheme rshift := make(Scheme, 0, maxMatch) rshift = append(rshift, s[0]) // 2. Iterate over matchers for p, i, l := 0, 1, len(s); i < l; i++ { pre, cur := s[p], s[i] /* Merge: 2 following literals */ if len(pre.pat) > 0 && len(cur.pat) > 0 { // merge strings into previous pre.pat = fmt.Sprintf("%s%s", pre.pat, cur.pat) // delete current s[i] = nil } // increment previous (only if current is not nul) if s[i] != nil { rshift = append(rshift, s[i]) p = i } } return rshift, nil } // matchString checks the STRING matchers from an URI // it returns a boolean : false when not matching, true eitherway // it returns a cleared uri, without STRING data func (s Scheme) matchString(uri string) (string, bool) { // 1. Initialise variables clr := uri // contains cleared input string minOff := 0 // minimum offset // 2. Iterate over strings for _, m := range s { ls := len(m.pat) // {1} If not STRING matcher -> ignore // if ls == 0 { continue } // {2} Get offset in URI (else -1) // off := strings.Index(clr, m.pat) if off < 0 { return "", false } // {3} Fail on invalid offset range // if off < minOff { return "", false } // {4} Check for trailing '/' // hasSlash := 0 if off+ls < len(clr) && clr[off+ls] == '/' { hasSlash = 1 } // {5} Remove the current string (+trailing slash) from the URI // beg, end := clr[:off], clr[off+ls+hasSlash:] clr = fmt.Sprintf("%s\a/%s", beg, end) // separate matches by '\a' character // {6} Update offset range // minOff = len(beg) + 2 - 1 // +2 slash separators // -1 because strings begin with 1 slash already } // 3. If exists, remove trailing '/' if clr[len(clr)-1] == '/' { clr = clr[:len(clr)-1] } // 4. If exists, remove trailing '\a' if clr[len(clr)-1] == '\a' { clr = clr[:len(clr)-1] } return clr, true } // matchWildcards check the WILCARDS (non-string) matchers from // a cleared URI. it returns if the string matches // + it sets the matchers buffers for later extraction func (s Scheme) matchWildcards(clear string) bool { // 1. Extract wildcards (ref) wildcards := make(Scheme, 0, maxMatch) for _, m := range s { if len(m.pat) == 0 { m.buf = nil // flush buffers wildcards = append(wildcards, m) } } // 2. If no wildcards -> match if len(wildcards) == 0 { return true } // 3. Break uri by '\a' characters matches := strings.Split(clear, "\a")[1:] // 4. Iterate over matches for n, match := range matches { // {1} If no more matcher // if n >= len(wildcards) { return false } // {2} Split by '/' // data := strings.Split(match, "/")[1:] // from index 1 because it begins with '/' // {3} If required and missing // if wildcards[n].req && len(data) < 1 { return false } // {4} If not multi but got multi // if !wildcards[n].mul && len(data) > 1 { return false } // {5} Store data into matcher // wildcards[n].buf = data } // 5. Match return true }