Faster key-value parsing

[gorecfile.git] / r.go

diff --git a/r.go b/r.go
index 36141715f151f5b8027ac962adf692897b40b4df..dc6084c3df116ca996f4c1d8709f80f58eeb15cf 100644 (file)
--- a/r.go
+++ b/r.go
@@ -1,6 +1,6 @@
 /*
 recfile -- GNU recutils'es recfiles parser on pure Go
-Copyright (C) 2020 Sergey Matveev <stargrave@stargrave.org>
+Copyright (C) 2020-2022 Sergey Matveev <stargrave@stargrave.org>
 
 This program is free software: you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
@@ -21,7 +21,6 @@ import (
        "bufio"
        "errors"
        "io"
-       "regexp"
        "strings"
 )
 
@@ -29,17 +28,39 @@ type Reader struct {
        scanner *bufio.Scanner
 }
 
+// Create Reader for iterating through the records. It uses
+// bufio.Scanner, so can read more than currently parsed records take.
 func NewReader(r io.Reader) *Reader {
        return &Reader{bufio.NewScanner(r)}
 }
 
-type Field struct {
-       Name  string
-       Value string
+func getKeyValue(text string) (string, string) {
+       cols := strings.SplitN(text, ":", 2)
+       if len(cols) != 2 {
+               return "", ""
+       }
+       k := cols[0]
+       if len(k) == 0 {
+               return "", ""
+       }
+       if !((k[0] == '%') ||
+               ('a' <= k[0] && k[0] <= 'z') ||
+               ('A' <= k[0] && k[0] <= 'Z')) {
+               return "", ""
+       }
+       for _, c := range k {
+               if !((c == '_') ||
+                       ('a' <= c && c <= 'z') ||
+                       ('A' <= c && c <= 'Z') ||
+                       ('0' <= c && c <= '9')) {
+                       return "", ""
+               }
+       }
+       return k, strings.TrimPrefix(cols[1], " ")
 }
 
-var KeyValRe = regexp.MustCompile(`([a-zA-Z%][a-zA-Z0-9_]*):\s*(.*)$`)
-
+// Get next record. Each record is just a collection of fields. io.EOF
+// is returned if there is nothing to read more.
 func (r *Reader) Next() ([]Field, error) {
        fields := make([]Field, 0, 1)
        var text string
@@ -95,12 +116,10 @@ func (r *Reader) Next() ([]Field, error) {
                        break
                }
 
-               matches := KeyValRe.FindStringSubmatch(text)
-               if len(matches) == 0 {
+               name, line = getKeyValue(text)
+               if name == "" {
                        return fields, errors.New("invalid field format")
                }
-               name = matches[1]
-               line = matches[2]
 
                if len(line) > 0 && line[len(line)-1] == '\\' {
                        continuation = true
@@ -123,3 +142,29 @@ func (r *Reader) Next() ([]Field, error) {
        }
        return fields, nil
 }
+
+// Same as Next(), but with unique keys and last value.
+func (r *Reader) NextMap() (map[string]string, error) {
+       fields, err := r.Next()
+       if err != nil {
+               return nil, err
+       }
+       m := make(map[string]string, len(fields))
+       for _, f := range fields {
+               m[f.Name] = f.Value
+       }
+       return m, nil
+}
+
+// Same as Next(), but with unique keys and slices of values.
+func (r *Reader) NextMapWithSlice() (map[string][]string, error) {
+       fields, err := r.Next()
+       if err != nil {
+               return nil, err
+       }
+       m := make(map[string][]string)
+       for _, f := range fields {
+               m[f.Name] = append(m[f.Name], f.Value)
+       }
+       return m, nil
+}