我希望提供一个结构化的配置文件,它对于非技术用户来说尽可能容易编辑(不幸的是它必须是一个文件),所以我想使用YAML。然而,我找不到任何方法从Unix shell脚本解析这个。


当前回答

使用Python的PyYAML或YAML::Perl等库最容易进行复杂的解析。

如果您希望将所有YAML值解析为bash值,请尝试此脚本。这也可以处理注释。参见下面的示例用法:

# pparse.py

import yaml
import sys
            
def parse_yaml(yml, name=''):
    if isinstance(yml, list):
        for data in yml:
            parse_yaml(data, name)
    elif isinstance(yml, dict):
        if (len(yml) == 1) and not isinstance(yml[list(yml.keys())[0]], list):
            print(str(name+'_'+list(yml.keys())[0]+'='+str(yml[list(yml.keys())[0]]))[1:])
        else:
            for key in yml:
                parse_yaml(yml[key], name+'_'+key)

            
if __name__=="__main__":
    yml = yaml.safe_load(open(sys.argv[1]))
    parse_yaml(yml)

test.yml

- folders:
  - temp_folder: datasets/outputs/tmp
  - keep_temp_folder: false

- MFA:
  - MFA: false
  - speaker_count: 1
  - G2P: 
    - G2P: true
    - G2P_model: models/MFA/G2P/english_g2p.zip
    - input_folder: datasets/outputs/Youtube/ljspeech/wavs
    - output_dictionary: datasets/outputs/Youtube/ljspeech/dictionary.dict
  - dictionary: datasets/outputs/Youtube/ljspeech/dictionary.dict
  - acoustic_model: models/MFA/acoustic/english.zip
  - temp_folder: datasets/outputs/tmp
  - jobs: 4
  - align:
    - config: configs/MFA/align.yaml
    - dataset: datasets/outputs/Youtube/ljspeech/wavs
    - output_folder: datasets/outputs/Youtube/ljspeech-aligned

- TTS:
  - output_folder: datasets/outputs/Youtube
  - preprocess:
    - preprocess: true
    - config: configs/TTS_preprocess.yaml # Default Config 
    - textgrid_folder: datasets/outputs/Youtube/ljspeech-aligned
    - output_duration_folder: datasets/outputs/Youtube/durations
    - sampling_rate: 44000 # Make sure sampling rate is same here as in preprocess config

需要YAML值的脚本:

yaml() {
    eval $(python pparse.py "$1")
}

yaml "test.yml"

# What python printed to bash:

folders_temp_folder=datasets/outputs/tmp
folders_keep_temp_folder=False
MFA_MFA=False
MFA_speaker_count=1
MFA_G2P_G2P=True
MFA_G2P_G2P_model=models/MFA/G2P/english_g2p.zip
MFA_G2P_input_folder=datasets/outputs/Youtube/ljspeech/wavs
MFA_G2P_output_dictionary=datasets/outputs/Youtube/ljspeech/dictionary.dict
MFA_dictionary=datasets/outputs/Youtube/ljspeech/dictionary.dict
MFA_acoustic_model=models/MFA/acoustic/english.zip
MFA_temp_folder=datasets/outputs/tmp
MFA_jobs=4
MFA_align_config=configs/MFA/align.yaml
MFA_align_dataset=datasets/outputs/Youtube/ljspeech/wavs
MFA_align_output_folder=datasets/outputs/Youtube/ljspeech-aligned
TTS_output_folder=datasets/outputs/Youtube
TTS_preprocess_preprocess=True
TTS_preprocess_config=configs/TTS_preprocess.yaml
TTS_preprocess_textgrid_folder=datasets/outputs/Youtube/ljspeech-aligned
TTS_preprocess_output_duration_folder=datasets/outputs/Youtube/durations
TTS_preprocess_sampling_rate=44000

使用bash访问变量:

echo "$TTS_preprocess_sampling_rate";
>>> 44000

其他回答

我的用例可能与这篇原始文章所要求的完全相同,也可能不完全相同,但它肯定是相似的。

我需要拉一些YAML作为bash变量。YAML的深度永远不会超过一层。

YAML看起来是这样的:

KEY:                value
ANOTHER_KEY:        another_value
OH_MY_SO_MANY_KEYS: yet_another_value
LAST_KEY:           last_value

输出如下:

KEY="value"
ANOTHER_KEY="another_value"
OH_MY_SO_MANY_KEYS="yet_another_value"
LAST_KEY="last_value"

我用这一行实现了输出:

sed -e 's/:[^:\/\/]/="/g;s/$/"/g;s/ *=/=/g' file.yaml > file.sh

s/:[^:\/\/]/="/g查找:并将其替换为=",同时忽略://(对于url) S /$/"/g将"附加到每一行的末尾 S / *=/=/g删除=前面的所有空格

下面是一个bash-only解析器,利用sed和awk来解析简单的yaml文件:

function parse_yaml {
   local prefix=$2
   local s='[[:space:]]*' w='[a-zA-Z0-9_]*' fs=$(echo @|tr @ '\034')
   sed -ne "s|^\($s\):|\1|" \
        -e "s|^\($s\)\($w\)$s:$s[\"']\(.*\)[\"']$s\$|\1$fs\2$fs\3|p" \
        -e "s|^\($s\)\($w\)$s:$s\(.*\)$s\$|\1$fs\2$fs\3|p"  $1 |
   awk -F$fs '{
      indent = length($1)/2;
      vname[indent] = $2;
      for (i in vname) {if (i > indent) {delete vname[i]}}
      if (length($3) > 0) {
         vn=""; for (i=0; i<indent; i++) {vn=(vn)(vname[i])("_")}
         printf("%s%s%s=\"%s\"\n", "'$prefix'",vn, $2, $3);
      }
   }'
}

它可以理解以下文件:

## global definitions
global:
  debug: yes
  verbose: no
  debugging:
    detailed: no
    header: "debugging started"

## output
output:
   file: "yes"

在解析时使用:

parse_yaml sample.yml

将输出:

global_debug="yes"
global_verbose="no"
global_debugging_detailed="no"
global_debugging_header="debugging started"
output_file="yes"

它也理解由ruby生成的yaml文件,其中可能包含ruby符号,例如:

---
:global:
  :debug: 'yes'
  :verbose: 'no'
  :debugging:
    :detailed: 'no'
    :header: debugging started
  :output: 'yes'

并将输出与前一个示例相同的结果。

脚本中的典型用法是:

eval $(parse_yaml sample.yml)

Parse_yaml接受一个前缀参数,这样导入的所有设置都有一个公共前缀(这将减少名称空间冲突的风险)。

parse_yaml sample.yml "CONF_"

收益率:

CONF_global_debug="yes"
CONF_global_verbose="no"
CONF_global_debugging_detailed="no"
CONF_global_debugging_header="debugging started"
CONF_output_file="yes"

注意,之前文件中的设置可以被后面的设置引用:

## global definitions
global:
  debug: yes
  verbose: no
  debugging:
    detailed: no
    header: "debugging started"

## output
output:
   debug: $global_debug

另一个很好的用法是先解析默认文件,然后解析用户设置,这是可行的,因为后一个设置会覆盖第一个设置:

eval $(parse_yaml defaults.yml)
eval $(parse_yaml project.yml)

你可以用golang写成yq的等价形式:

./go-yg -yamlFile /home/user/dev/ansible-firefox/defaults/main.yml -key
firefox_version

返回:

62.0.3

我已经用python编写了shyaml,用于从shell命令行查询YAML。

概述:

$ pip install shyaml      ## installation

示例的YAML文件(具有复杂的功能):

$ cat <<EOF > test.yaml
name: "MyName !!"
subvalue:
    how-much: 1.1
    things:
        - first
        - second
        - third
    other-things: [a, b, c]
    maintainer: "Valentin Lab"
    description: |
        Multiline description:
        Line 1
        Line 2
EOF

基本的查询:

$ cat test.yaml | shyaml get-value subvalue.maintainer
Valentin Lab

更复杂的循环查询复杂的值:

$ cat test.yaml | shyaml values-0 | \
  while read -r -d $'\0' value; do
      echo "RECEIVED: '$value'"
  done
RECEIVED: '1.1'
RECEIVED: '- first
- second
- third'
RECEIVED: '2'
RECEIVED: 'Valentin Lab'
RECEIVED: 'Multiline description:
Line 1
Line 2'

以下几个要点:

all YAML types and syntax oddities are correctly handled, as multiline, quoted strings, inline sequences... \0 padded output is available for solid multiline entry manipulation. simple dotted notation to select sub-values (ie: subvalue.maintainer is a valid key). access by index is provided to sequences (ie: subvalue.things.-1 is the last element of the subvalue.things sequence.) access to all sequence/structs elements in one go for use in bash loops. you can output whole subpart of a YAML file as ... YAML, which blend well for further manipulations with shyaml.

更多的示例和文档可以在shyaml github页面或shyaml PyPI页面上找到。

Whenever you need a solution for "How to work with YAML/JSON/compatible data from a shell script" which works on just about every OS with Python (*nix, OSX, Windows), consider yamlpath, which provides several command-line tools for reading, writing, searching, and merging YAML, EYAML, JSON, and compatible files. Since just about every OS either comes with Python pre-installed or it is trivial to install, this makes yamlpath highly portable. Even more interesting: this project defines an intuitive path language with very powerful, command-line-friendly syntax that enables accessing one or more nodes.

针对您的具体问题,在使用Python的本地包管理器或您的操作系统的包管理器安装yamlpath之后(yamlpath可以通过RPM对某些操作系统提供):

#!/bin/bash
# Read values directly from YAML (or EYAML, JSON, etc) for use in this shell script:
myShellVar=$(yaml-get --query=any.path.no[matter%how].complex source-file.yaml)

# Use the value any way you need:
echo "Retrieved ${myShellVar}"

# Perhaps change the value and write it back:
myShellVar="New Value"
yaml-set --change=/any/path/no[matter%how]/complex --value="$myShellVar" source-file.yaml

不过,您没有指定数据是一个简单的Scalar值,因此让我们提高赌注。如果你想要的结果是一个数组呢?更有挑战性的是,如果它是一个哈希数组,而你只想要每个结果的一个属性呢?进一步假设您的数据实际上分布在多个YAML文件中,并且您需要在单个查询中获得所有结果。这是一个更有趣的问题。所以,假设你有这两个YAML文件:

文件:data1.yaml

---
baubles:
  - name: Doohickey
    sku: 0-000-1
    price: 4.75
    weight: 2.7g
  - name: Doodad
    sku: 0-000-2
    price: 10.5
    weight: 5g
  - name: Oddball
    sku: 0-000-3
    price: 25.99
    weight: 25kg

文件:data2.yaml

---
baubles:
  - name: Fob
    sku: 0-000-4
    price: 0.99
    weight: 18mg
  - name: Doohickey
    price: 10.5
  - name: Oddball
    sku: 0-000-3
    description: This ball is odd

在应用数据2的更改后,如何仅报告库存中每个项目的sku。Yaml到data1。Yaml,所有从一个shell脚本?试试这个:

#!/bin/bash
baubleSKUs=($(yaml-merge --aoh=deep data1.yaml data2.yaml | yaml-get --query=/baubles/sku -))

for sku in "${baubleSKUs[@]}"; do
    echo "Found bauble SKU:  ${sku}"
done

你只需要几行代码就能得到你想要的东西:

Found bauble SKU:  0-000-1
Found bauble SKU:  0-000-2
Found bauble SKU:  0-000-3
Found bauble SKU:  0-000-4

如您所见,yamlpath将非常复杂的问题转化为简单的解决方案。注意,整个查询是作为一个流处理的;查询没有更改YAML文件,也没有临时文件。

I realize this is "yet another tool to solve the same question" but after reading the other answers here, yamlpath appears more portable and robust than most alternatives. It also fully understands YAML/JSON/compatible files and it does not need to convert YAML to JSON to perform requested operations. As such, comments within the original YAML file are preserved whenever you need to change data in the source YAML file. Like some alternatives, yamlpath is also portable across OSes. More importantly, yamlpath defines a query language that is extremely powerful, enabling very specialized/filtered data queries. It can even operate against results from disparate parts of the file in a single query.

If you want to get or set many values in the data at once -- including complex data like hashes/arrays/maps/lists -- yamlpath can do that. Want a value but don't know precisely where it is in the document? yamlpath can find it and give you the exact path(s). Need to merge multiple data file together, including from STDIN? yamlpath does that, too. Further, yamlpath fully comprehends YAML anchors and their aliases, always giving or changing exactly the data you expect whether it is a concrete or referenced value.

免责声明:我编写并维护了yamlpath,它是基于ruamel的。yaml是基于PyYAML的。因此,yamlpath完全符合标准。