Documentation
¶
Overview ¶
Package utils defines parsing and run command functions that can be used outside nodeprofiler.
Index ¶
- func ParseColumns(rows []string, allTitles []string, wantTitles ...string) (map[string][]string, error)
- func ParseRows(lines []string, delim string, titles ...string) (map[string][]string, error)
- func ParseRowsAndColumns(lines []string, titles ...string) (map[string][]string, error)
- func RunCommand(cmd string, args ...string) ([]byte, error)
- func SumAtoi(a []string) (int, error)
- func SumParseFloat(a []string) (float64, error)
- func TrimCharacter(a []string, char string) []string
- type ParsedOutput
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func ParseColumns ¶
func ParseColumns(rows []string, allTitles []string, wantTitles ...string) (map[string][]string, error)
ParseColumns parses command outputs which are in a column table. It takes in a slice of strings which has all the column titles of the command in the correct order. This is necessary because some titles contain multiple strings, thus splitting by whitespaces will split row incorrectly. E.g, splitting the titles row in df's output based on whitespaces will result in:
"Filesystem Use% Mounted on" -> ["Filesystem", "Use%", "Mounted", "on"] instead of:
["Filesystem", "Use%", "Mounted on"]
The allTitles slice is used to give each title its correct index. If the titles in the slice are in wrong order, the function's output will be incorrect.
The function also takes in an optional want titles slice which specifies the columns to parse. If this argument is missing, then all columns are parsed.
Eg, ParseColumns(["r b swpd buff",
"10 0 14831128 0"],
["r", "b"]) = map[string][]string {
"r": ["10"]
"b": ["0"]}
The output needs to have titles on all its columns else the function will return an error:
Eg [FAIL] ParseColumns([" total used",
"Mem: 14520 12", "Swap: 0 0"], ["total", "used"]) err : "row has different number of columns from header row"
Some edge cases that will be parsed by this function include:
rows with repeated headers, eg with the output of iostat: [] string {"Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn",
"vdb 2.39 57.39 69.83 855576 1041132", " ", "Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn", " ", "Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn"} =>
map[string][]string {
"tps" : {"2.39"},
"kB_read/s" : {"57.39"},
...
}
func ParseRows ¶
ParseRows parses command outputs which are in a row table. It takes in a string which specifies the delimiter that separates row title from values. The function does not support '\n' as a delimiter for now. It takes in an optional titles slice which specifies which rows to parse.
Eg, ParseRows(["avg-cpu: %user %nice %system %iowait %steal %idle"],
[avg-cpu]) = map[string][]string {
avg-cpu: [%user, %nice, %system, %iowait, %steal, %idle]}
If the wrong delimiter is passed in, the function returns an error:
Eg [FAIL] ParseRows(["Device tps kB_read/s kB_wrtn/s",
"vdb 1.13 19.48 33.61" "vda 0.02 0.86 0.00"], ":", ["vda"]) err: "failed to split row into row title and value"
Some edge cases parsed by this function include:
Rows whose delimiter have whitespaces around it. For example, [] string { "processor: 7",
"CPU family: 6"} =>
map[string][]string {
"processor" : {"7"}
"cpu family" : {"6"}
}
OR ¶
[] string { "processor : 7",
"cpu family : 6"} =>
map[string][]string {
"processor" : {"7"}
"cpu family" : {"6"}
}
func ParseRowsAndColumns ¶
ParseRowsAndColumns parses command outputs that have row and column headers. Specifically, it parses command outputs that have an empty string on row 0 column 0. For example, the output of free as seen below:
total used free
Mem: 14518 12 14479 Swap: 0 0 0
If the command output has a non-empty string on row 0 column 0 as with iostat below, the function will throw an error since this is an output that should be parsed by columns instead "Device tps kB_read/s kB_wrtn/s kB_read kB_wrtn", "vdb 0.74 10.39 23.23 859900 1922916", "vda 0.01 0.46 0.00 37829 0"
The function takes in an optional titles slice which specifies the row column combination to parse. If the titles argument is missing, then an empty map is returned.
Eg, ParseRowsAndColumns([" total used free shared",
"Mem: 14520 12 14482 0",
"Swap: 0 0 0 "],
["Mem:used", "Swap:total"]) = map[string][]string {
"Mem:used": ["12"]
"Swap:total" : ["0"]}
The titles should be in the format "row:column". Else, an error is returned:
Eg [FAIL], ParseRowsAndColumns([" total used free shared",
"Mem: 14520 12 14482 0", "Swap: 0 0 0 "], ["Mem+used", "Swap+total"]) err : "title string not well-formatted"
func RunCommand ¶
RunCommand is a wrapper function for exec.Command that will run the command specified return its output and/or error.
func SumAtoi ¶
SumAtoi converts all the strings in a slice to integers, sums them up and returns the result. A non-nil error is returned if an error occurred.
func SumParseFloat ¶
SumParseFloat converts all the strings in a slice to floating points, sums them up, and returns the result as a floating point. A non-nil error is returned if an error occurred.
func TrimCharacter ¶
TrimCharacter trims the specified character from each string in a slice of strings and returns a slice with the trimmed strings.
Types ¶
type ParsedOutput ¶
ParsedOutput is a data structure that holds the parsed output of certain shell commands whose output takes the form of a table.
Source Files