解析io.ReadAll以匹配特定的表达式

Question

I'm using the Golang Docker SDK to output container logs. The container is running a scan and outputs specific information about the scan job start time, end time, mean scan duration as below:

    Selected XML parser javax.xml.bind.util.JAXBSource$1 does not recognize the feature http://xml.org/sax/features/validation
    Generated ./reports/CSR1000V_RTR2.json
    Generated ./reports/CSR1000V_RTR6.json
    Generated ./reports/CSR1000V_RTR3.json
Scan start time: Mon Aug 27 03:39:24 GMT 2018
Scan end time:   Mon Aug 27 03:39:40 GMT 2018
Mean target scan duration: 3906ms

I use code below to convert the io.Reader into a string:

out, err := cli.ContainerLogs(ctx, resp.ID, types.ContainerLogsOptions{
    ShowStdout: true,
    Follow:     true})
if err != nil {
    panic(err)
}
defer out.Close()
//io.Copy(os.Stdout, out)
b, err := ioutil.ReadAll(out)

fmt.Println(string(b))

How in Golang can I parse only the last 3 lines and capture the following values only from the stdout:

fmt.Println("The scan has started at: " +startime)
fmt.Println("The scan has ended at: " +endtime)
fmt.Println("The scan job took xxx ms to scan each device")

Answer 1

An alternative solution is to use the bufio.Scanner api.

It might be interesting if you are dealing with large output that you don t want to store full in memory for processing.

The scanner is a composition of a reader, the input data, and a split function, to creates meaningful chunks of data.

Using the standard api it is possible to take advantage of the provided bufio.ScanLines function to split the output by lines, then a simple prefix equality will yield the necessary information to identify the researched information.

package main

import (
    "bufio"
    "bytes"
    "fmt"
    "strings"
)

type result struct {
    start string
    end   string
    mean  string
}

func main() {
    raw := `
    Selected XML parser javax.xml.bind.util.JAXBSource$1 does not recognize the feature http://xml.org/sax/features/validation
    Generated ./reports/CSR1000V_RTR2.json
    Generated ./reports/CSR1000V_RTR6.json
    Generated ./reports/CSR1000V_RTR3.json
Scan start time: Mon Aug 27 03:39:24 GMT 2018
Scan end time:   Mon Aug 27 03:39:40 GMT 2018
Mean target scan duration: 3906ms
`
    scanner := bufio.NewScanner(strings.NewReader(raw))

    var res result
    // Create a custom split function by wrapping the existing ScanLines function.
    split := func(data []byte, atEOF bool) (advance int, token []byte, err error) {
        advance, token, err = bufio.ScanLines(data, atEOF)
        if err == nil && token != nil {
            if pattern := []byte("Scan start time:"); bytes.HasPrefix(token, pattern) {
                res.start = strings.TrimSpace(string(token[len(pattern):]))
            } else if pattern := []byte("Scan end time:"); bytes.HasPrefix(token, pattern) {
                res.end = strings.TrimSpace(string(token[len(pattern):]))
            } else if pattern := []byte("Mean target scan duration:"); bytes.HasPrefix(token, pattern) {
                res.mean = strings.TrimSpace(string(token[len(pattern):]))
            }
        }
        return
    }
    // Set the split function for the scanning operation.
    scanner.Split(split)
    // drain the source
    for scanner.Scan() {
    }

    if err := scanner.Err(); err != nil {
        fmt.Printf("Invalid input: %s", err)
    }
    fmt.Printf("%#v
", res)
}