I'm using ruamel.yaml to generate a YAML file that will be read by Tavern, which requires the file to contain a list like this:
includes:
- !include vars.yaml
Attempting to use any of the usual approaches to dump the data as strings results in single quotes being added around the tags, which doesn't work when the YAML is ingested by the next tool.
How do I generate a YAML file that contains unquoted local tags, starting with data that is defined in a dictionary?
I was able to create a YAML file with the required format using the following approach, based on prior examples. My approach is more flexible because it allows the tag handle to be an instance property rather than a class property, so you don't need to define a different class for every tag handle.
import sys
from ruamel.yaml import YAML
yaml = YAML(typ='rt')
class TaggedString:
def __init__(self, handle, value):
self.handle = handle
self.value = value
#classmethod
def to_yaml(cls, representer, node):
# I don't understand the arguments to the following function!
return representer.represent_scalar(u'{.handle}'.format(node),
u'{.value}'.format(node))
yaml.register_class(TaggedString)
data = {
'includes': [
TaggedString('!include', 'vars.yaml'),
TaggedString('!exclude', 'dummy.yaml')
]
}
yaml.dump(data, sys.stdout)
Output:
includes:
- !include vars.yaml
- !exclude dummy.yaml
I am not sure if this is the best approach. I might be missing a simpler way to achieve the same result. Note that my goal is not to dump a Python class; I'm just doing that as a way to get the tag to be written correctly.
I am not sure if this is a better approach, but if you had tried to round-trip your required output, you
would have seen that ruamel.yaml actually can preserve your tagged strings, without you having to
do anything. Inspecting the Python datastructure, you'll notice that ruamel.yaml does
this by creating a TaggedScalar (as you cannnot attach attributes to the built-in string type).
import sys
import ruamel.yaml
yaml_str = """\
includes:
- !include vars.yaml
- !exclude dummy.yaml
"""
yaml = ruamel.yaml.YAML()
data = yaml.load(yaml_str)
yaml.dump(data, sys.stdout)
incl = data['includes'][0]
print(type(incl))
which gives:
includes:
- !include vars.yaml
- !exclude dummy.yaml
<class 'ruamel.yaml.comments.TaggedScalar'>
After inspecting comments.py (and possible constructor.py), you should be able
to make ruamel.yaml's internal data structure on the fly:
import sys
import ruamel.yaml
from ruamel.yaml.comments import TaggedScalar
def tagged_string(tag, val):
# starting with ruamel.yaml>0.16.5 you can replace the following lines with:
# return TaggedScalar(value=val, tag=tag)
ret_val = TaggedScalar()
ret_val.value = val
ret_val.yaml_set_tag(tag)
return ret_val
yaml = ruamel.yaml.YAML()
data = dict(includes=[tagged_string('!include', 'vars.yaml'),
tagged_string('!include', 'vars.yaml'),
])
yaml.dump(data, sys.stdout)
which also gives:
includes:
- !include vars.yaml
- !include vars.yaml
Let's say I have a (simplified) class as below. I am using it for a program configuration (hyperparameters).
# config.py
class Config(object): # default configuration
GPU_COUNT = 1
IMAGES_PER_GPU = 2
MAP = {1:2, 2:3}
def display(self):
pass
# experiment1.py
from config import Config as Default
class Config(Default): # some over-written configuration
GPU_COUNT = 2
NAME='2'
# run.py
from experiment1 import Config
cfg = Config()
...
cfg.NAME = 'ABC' # possible runtime over-writing
# Now I would like to save `cfg` at this moment
I'd like to save this configuration and restore later. The member functions must be out of concern when restoring.
1. When I tried pickle:
import pickle
with open('cfg.pk', 'rb') as f: cfg = pickle.load(f)
##--> AttributeError: Can't get attribute 'Config' on <module '__main__'>
I saw a solution using class_def of Config, but I wish I can restore the configuration without knowing the class definition (eg, export to dict and save as JSON)
2. I tried to convert class to dict (so that I can export as JSON)
cfg.__dict__ # {'NAME': 'ABC'}
vars(cfg) # {'NAME': 'ABC'}
In both cases, it was difficult to access attributes. Is it possible?
The question's title is "how to convert python class to dict", but I suspect you are really just looking for an easy way to represent (hyper)parameters.
By far the easiest solution is to not use classes for this. I've seen it happen on some machine learning tutorials, but I consider it a pretty ugly hack. It breaks some semantics about classes vs objects, and the difficulty pickling is a result from that. How about you use a simple class like this one:
class Params(dict):
__getattr__ = dict.__getitem__
__setattr__ = dict.__setitem__
__delattr__ = dict.__delitem__
def __getstate__(self):
return self
def __setstate__(self, state):
self.update(state)
def copy(self, **extra_params):
return Params(**self, **extra_params)
It can do everything the class approach can. Predefined configs are then just objects you should copy before editing, as follows:
config = Params(
GPU_COUNT = 2,
NAME='2',
)
other_config = config.copy()
other_config.GPU_COUNT = 4
Or alternatively in one step:
other_config = config.copy(
GPU_COUNT = 4
)
Works fine with pickle (although you will need to have the Params class somewhere in your source), and you could also easily write load and save methods for the Params class if you want to use JSON.
In short, do not use a class for something that really is just an object.
Thankfully, #evertheylen's answer was great to me. However, the code returns error when p.__class__ = Params, so I slightly changed as below. I think it works in the same way.
class Params(dict):
__getattr__ = dict.__getitem__
__setattr__ = dict.__setitem__
__delattr__ = dict.__delitem__
def __getstate__(self):
return self
def __setstate__(self, state):
self.update(state)
def copy(self, **extra_params):
lhs = Params()
lhs.update(self)
lhs.update(extra_params)
return lhs
and you can do
config = Params(
GPU_COUNT = 2,
NAME='2',
)
other_config = config.copy()
other_config.GPU_COUNT = 4
I am new to Python. I have to edit some test I have.
I have a file which inclused all the details for a test I have to execute. The file name for example is "test 1"
In this file there are few global params:
LINK_IP_ADDR = "10.8.22.89"
STABILITY_SLEEP = 5
MAIN_STATE_SLEEP = 30
For making the tests automaticly and generic I want to set the the value from the params from another python file (for example config_file). How should I do it?
You should probably use configuration files for this. The ConfigParser package would do the job.
If you want a quick and dirty version, the following could be for you.
In your 'test files', write the following:
class Params:
LINK_IP_ADDR = "10.8.22.89"
STABILITY_SLEEP = 5
MAIN_STATE_SLEEP = 30
Then in your main program you import the parameters you want and just use them:
from test1 import Params
# from test2 import Params
print Params.LINK_IP_ADDR
You can then switch your import statements to the file you want to use. I would use a normal .ini file with different sections though.
Here is a quick n' dirty way from a plain file:
params = dict()
with open('config_file') as f:
for line in f:
param, value = line.split("=")
params[param.strip()] = value.strip()
print(params)
config_file:
LINK_IP_ADDR = "10.8.22.89"
STABILITY_SLEEP = 5
MAIN_STATE_SLEEP = 30
for config file i used .yaml file. This format stores your config like json or python-dict. U can store complex structure (array, hashmap, ...) very good
config.yaml
LINK_IP_ADDR: 10.8.22.89
STABILITY_SLEEP: 5
MAIN_STATE_SLEEP: 30
another.py
import yaml
configs = yaml.safe_load(file("path_to_config.yaml"))
# to use you configs:
configs['LINK_IP_ADDR']
#.....
Given the following format (.properties or .ini):
propertyName1=propertyValue1
propertyName2=propertyValue2
...
propertyNameN=propertyValueN
For Java there is the Properties class that offers functionality to parse / interact with the above format.
Is there something similar in python's standard library (2.x) ?
If not, what other alternatives do I have ?
I was able to get this to work with ConfigParser, no one showed any examples on how to do this, so here is a simple python reader of a property file and example of the property file. Note that the extension is still .properties, but I had to add a section header similar to what you see in .ini files... a bit of a bastardization, but it works.
The python file: PythonPropertyReader.py
#!/usr/bin/python
import ConfigParser
config = ConfigParser.RawConfigParser()
config.read('ConfigFile.properties')
print config.get('DatabaseSection', 'database.dbname');
The property file: ConfigFile.properties
[DatabaseSection]
database.dbname=unitTest
database.user=root
database.password=
For more functionality, read: https://docs.python.org/2/library/configparser.html
For .ini files there is the configparser module that provides a format compatible with .ini files.
Anyway there's nothing available for parsing complete .properties files, when I have to do that I simply use jython (I'm talking about scripting).
I know that this is a very old question, but I need it just now and I decided to implement my own solution, a pure python solution, that covers most uses cases (not all):
def load_properties(filepath, sep='=', comment_char='#'):
"""
Read the file passed as parameter as a properties file.
"""
props = {}
with open(filepath, "rt") as f:
for line in f:
l = line.strip()
if l and not l.startswith(comment_char):
key_value = l.split(sep)
key = key_value[0].strip()
value = sep.join(key_value[1:]).strip().strip('"')
props[key] = value
return props
You can change the sep to ':' to parse files with format:
key : value
The code parses correctly lines like:
url = "http://my-host.com"
name = Paul = Pablo
# This comment line will be ignored
You'll get a dict with:
{"url": "http://my-host.com", "name": "Paul = Pablo" }
A java properties file is often valid python code as well. You could rename your myconfig.properties file to myconfig.py. Then just import your file, like this
import myconfig
and access the properties directly
print myconfig.propertyName1
if you don't have multi line properties and a very simple need, a few lines of code can solve it for you:
File t.properties:
a=b
c=d
e=f
Python code:
with open("t.properties") as f:
l = [line.split("=") for line in f.readlines()]
d = {key.strip(): value.strip() for key, value in l}
If you have an option of file formats I suggest using .ini and Python's ConfigParser as mentioned. If you need compatibility with Java .properties files I have written a library for it called jprops. We were using pyjavaproperties, but after encountering various limitations I ended up implementing my own. It has full support for the .properties format, including unicode support and better support for escape sequences. Jprops can also parse any file-like object while pyjavaproperties only works with real files on disk.
This is not exactly properties but Python does have a nice library for parsing configuration files. Also see this recipe: A python replacement for java.util.Properties.
i have used this, this library is very useful
from pyjavaproperties import Properties
p = Properties()
p.load(open('test.properties'))
p.list()
print(p)
print(p.items())
print(p['name3'])
p['name3'] = 'changed = value'
Here is link to my project: https://sourceforge.net/projects/pyproperties/. It is a library with methods for working with *.properties files for Python 3.x.
But it is not based on java.util.Properties
This is a one-to-one replacement of java.util.Propeties
From the doc:
def __parse(self, lines):
""" Parse a list of lines and create
an internal property dictionary """
# Every line in the file must consist of either a comment
# or a key-value pair. A key-value pair is a line consisting
# of a key which is a combination of non-white space characters
# The separator character between key-value pairs is a '=',
# ':' or a whitespace character not including the newline.
# If the '=' or ':' characters are found, in the line, even
# keys containing whitespace chars are allowed.
# A line with only a key according to the rules above is also
# fine. In such case, the value is considered as the empty string.
# In order to include characters '=' or ':' in a key or value,
# they have to be properly escaped using the backslash character.
# Some examples of valid key-value pairs:
#
# key value
# key=value
# key:value
# key value1,value2,value3
# key value1,value2,value3 \
# value4, value5
# key
# This key= this value
# key = value1 value2 value3
# Any line that starts with a '#' is considerered a comment
# and skipped. Also any trailing or preceding whitespaces
# are removed from the key/value.
# This is a line parser. It parses the
# contents like by line.
You can use a file-like object in ConfigParser.RawConfigParser.readfp defined here -> https://docs.python.org/2/library/configparser.html#ConfigParser.RawConfigParser.readfp
Define a class that overrides readline that adds a section name before the actual contents of your properties file.
I've packaged it into the class that returns a dict of all the properties defined.
import ConfigParser
class PropertiesReader(object):
def __init__(self, properties_file_name):
self.name = properties_file_name
self.main_section = 'main'
# Add dummy section on top
self.lines = [ '[%s]\n' % self.main_section ]
with open(properties_file_name) as f:
self.lines.extend(f.readlines())
# This makes sure that iterator in readfp stops
self.lines.append('')
def readline(self):
return self.lines.pop(0)
def read_properties(self):
config = ConfigParser.RawConfigParser()
# Without next line the property names will be lowercased
config.optionxform = str
config.readfp(self)
return dict(config.items(self.main_section))
if __name__ == '__main__':
print PropertiesReader('/path/to/file.properties').read_properties()
If you need to read all values from a section in properties file in a simple manner:
Your config.properties file layout :
[SECTION_NAME]
key1 = value1
key2 = value2
You code:
import configparser
config = configparser.RawConfigParser()
config.read('path_to_config.properties file')
details_dict = dict(config.items('SECTION_NAME'))
This will give you a dictionary where keys are same as in config file and their corresponding values.
details_dict is :
{'key1':'value1', 'key2':'value2'}
Now to get key1's value :
details_dict['key1']
Putting it all in a method which reads that section from config file only once(the first time the method is called during a program run).
def get_config_dict():
if not hasattr(get_config_dict, 'config_dict'):
get_config_dict.config_dict = dict(config.items('SECTION_NAME'))
return get_config_dict.config_dict
Now call the above function and get the required key's value :
config_details = get_config_dict()
key_1_value = config_details['key1']
-------------------------------------------------------------
Extending the approach mentioned above, reading section by section automatically and then accessing by section name followed by key name.
def get_config_section():
if not hasattr(get_config_section, 'section_dict'):
get_config_section.section_dict = dict()
for section in config.sections():
get_config_section.section_dict[section] =
dict(config.items(section))
return get_config_section.section_dict
To access:
config_dict = get_config_section()
port = config_dict['DB']['port']
(here 'DB' is a section name in config file
and 'port' is a key under section 'DB'.)
create a dictionary in your python module and store everything into it and access it, for example:
dict = {
'portalPath' : 'www.xyx.com',
'elementID': 'submit'}
Now to access it you can simply do:
submitButton = driver.find_element_by_id(dict['elementID'])
My Java ini files didn't have section headers and I wanted a dict as a result. So i simply injected an "[ini]" section and let the default config library do its job.
Take a version.ini fie of the eclipse IDE .metadata directory as an example:
#Mon Dec 20 07:35:29 CET 2021
org.eclipse.core.runtime=2
org.eclipse.platform=4.19.0.v20210303-1800
# 'injected' ini section
[ini]
#Mon Dec 20 07:35:29 CET 2021
org.eclipse.core.runtime=2
org.eclipse.platform=4.19.0.v20210303-1800
The result is converted to a dict:
from configparser import ConfigParser
#staticmethod
def readPropertyFile(path):
# https://stackoverflow.com/questions/3595363/properties-file-in-python-similar-to-java-properties
config = ConfigParser()
s_config= open(path, 'r').read()
s_config="[ini]\n%s" % s_config
# https://stackoverflow.com/a/36841741/1497139
config.read_string(s_config)
items=config.items('ini')
itemDict={}
for key,value in items:
itemDict[key]=value
return itemDict
This is what I'm doing in my project: I just create another .py file called properties.py which includes all common variables/properties I used in the project, and in any file need to refer to these variables, put
from properties import *(or anything you need)
Used this method to keep svn peace when I was changing dev locations frequently and some common variables were quite relative to local environment. Works fine for me but not sure this method would be suggested for formal dev environment etc.
import json
f=open('test.json')
x=json.load(f)
f.close()
print(x)
Contents of test.json:
{"host": "127.0.0.1", "user": "jms"}
I have created a python module that is almost similar to the Properties class of Java ( Actually it is like the PropertyPlaceholderConfigurer in spring which lets you use ${variable-reference} to refer to already defined property )
EDIT : You may install this package by running the command(currently tested for python 3).
pip install property
The project is hosted on GitHub
Example : ( Detailed documentation can be found here )
Let's say you have the following properties defined in my_file.properties file
foo = I am awesome
bar = ${chocolate}-bar
chocolate = fudge
Code to load the above properties
from properties.p import Property
prop = Property()
# Simply load it into a dictionary
dic_prop = prop.load_property_files('my_file.properties')
Below 2 lines of code shows how to use Python List Comprehension to load 'java style' property file.
split_properties=[line.split("=") for line in open('/<path_to_property_file>)]
properties={key: value for key,value in split_properties }
Please have a look at below post for details
https://ilearnonlinesite.wordpress.com/2017/07/24/reading-property-file-in-python-using-comprehension-and-generators/
you can use parameter "fromfile_prefix_chars" with argparse to read from config file as below---
temp.py
parser = argparse.ArgumentParser(fromfile_prefix_chars='#')
parser.add_argument('--a')
parser.add_argument('--b')
args = parser.parse_args()
print(args.a)
print(args.b)
config file
--a
hello
--b
hello dear
Run command
python temp.py "#config"
You could use - https://pypi.org/project/property/
eg - my_file.properties
foo = I am awesome
bar = ${chocolate}-bar
chocolate = fudge
long = a very long property that is described in the property file which takes up \
multiple lines can be defined by the escape character as it is done here
url=example.com/api?auth_token=xyz
user_dir=${HOME}/test
unresolved = ${HOME}/files/${id}/${bar}/
fname_template = /opt/myapp/{arch}/ext/{objid}.dat
Code
from properties.p import Property
## set use_env to evaluate properties from shell / os environment
prop = Property(use_env = True)
dic_prop = prop.load_property_files('my_file.properties')
## Read multiple files
## dic_prop = prop.load_property_files('file1', 'file2')
print(dic_prop)
# Output
# {'foo': 'I am awesome', 'bar': 'fudge-bar', 'chocolate': 'fudge',
# 'long': 'a very long property that is described in the property file which takes up multiple lines
# can be defined by the escape character as it is done here', 'url': 'example.com/api?auth_token=xyz',
# 'user_dir': '/home/user/test',
# 'unresolved': '/home/user/files/${id}/fudge-bar/',
# 'fname_template': '/opt/myapp/{arch}/ext/{objid}.dat'}
I did this using ConfigParser as follows. The code assumes that there is a file called config.prop in the same directory where BaseTest is placed:
config.prop
[CredentialSection]
app.name=MyAppName
BaseTest.py:
import unittest
import ConfigParser
class BaseTest(unittest.TestCase):
def setUp(self):
__SECTION = 'CredentialSection'
config = ConfigParser.ConfigParser()
config.readfp(open('config.prop'))
self.__app_name = config.get(__SECTION, 'app.name')
def test1(self):
print self.__app_name % This should print: MyAppName
This is what i had written to parse file and set it as env variables which skips comments and non key value lines added switches to specify
hg:d
-h or --help print usage summary
-c Specify char that identifies comment
-s Separator between key and value in prop file
and specify properties file that needs to be parsed eg : python
EnvParamSet.py -c # -s = env.properties
import pipes
import sys , getopt
import os.path
class Parsing :
def __init__(self , seprator , commentChar , propFile):
self.seprator = seprator
self.commentChar = commentChar
self.propFile = propFile
def parseProp(self):
prop = open(self.propFile,'rU')
for line in prop :
if line.startswith(self.commentChar)==False and line.find(self.seprator) != -1 :
keyValue = line.split(self.seprator)
key = keyValue[0].strip()
value = keyValue[1].strip()
print("export %s=%s" % (str (key),pipes.quote(str(value))))
class EnvParamSet:
def main (argv):
seprator = '='
comment = '#'
if len(argv) is 0:
print "Please Specify properties file to be parsed "
sys.exit()
propFile=argv[-1]
try :
opts, args = getopt.getopt(argv, "hs:c:f:", ["help", "seprator=","comment=", "file="])
except getopt.GetoptError,e:
print str(e)
print " possible arguments -s <key value sperator > -c < comment char > <file> \n Try -h or --help "
sys.exit(2)
if os.path.isfile(args[0])==False:
print "File doesnt exist "
sys.exit()
for opt , arg in opts :
if opt in ("-h" , "--help"):
print " hg:d \n -h or --help print usage summary \n -c Specify char that idetifes comment \n -s Sperator between key and value in prop file \n specify file "
sys.exit()
elif opt in ("-s" , "--seprator"):
seprator = arg
elif opt in ("-c" , "--comment"):
comment = arg
p = Parsing( seprator, comment , propFile)
p.parseProp()
if __name__ == "__main__":
main(sys.argv[1:])
Lightbend has released the Typesafe Config library, which parses properties files and also some JSON-based extensions. Lightbend's library is only for the JVM, but it seems to be widely adopted and there are now ports in many languages, including Python: https://github.com/chimpler/pyhocon
You can use the following function, which is the modified code of #mvallebr. It respects the properties file comments, ignores empty new lines, and allows retrieving a single key value.
def getProperties(propertiesFile ="/home/memin/.config/customMemin/conf.properties", key=''):
"""
Reads a .properties file and returns the key value pairs as dictionary.
if key value is specified, then it will return its value alone.
"""
with open(propertiesFile) as f:
l = [line.strip().split("=") for line in f.readlines() if not line.startswith('#') and line.strip()]
d = {key.strip(): value.strip() for key, value in l}
if key:
return d[key]
else:
return d
this works for me.
from pyjavaproperties import Properties
p = Properties()
p.load(open('test.properties'))
p.list()
print p
print p.items()
print p['name3']
I followed configparser approach and it worked quite well for me. Created one PropertyReader file and used config parser there to ready property to corresponding to each section.
**Used Python 2.7
Content of PropertyReader.py file:
#!/usr/bin/python
import ConfigParser
class PropertyReader:
def readProperty(self, strSection, strKey):
config = ConfigParser.RawConfigParser()
config.read('ConfigFile.properties')
strValue = config.get(strSection,strKey);
print "Value captured for "+strKey+" :"+strValue
return strValue
Content of read schema file:
from PropertyReader import *
class ReadSchema:
print PropertyReader().readProperty('source1_section','source_name1')
print PropertyReader().readProperty('source2_section','sn2_sc1_tb')
Content of .properties file:
[source1_section]
source_name1:module1
sn1_schema:schema1,schema2,schema3
sn1_sc1_tb:employee,department,location
sn1_sc2_tb:student,college,country
[source2_section]
source_name1:module2
sn2_schema:schema4,schema5,schema6
sn2_sc1_tb:employee,department,location
sn2_sc2_tb:student,college,country
You can try the python-dotenv library. This library reads key-value pairs from a .env (so not exactly a .properties file though) file and can set them as environment variables.
Here's a sample usage from the official documentation:
from dotenv import load_dotenv
load_dotenv() # take environment variables from .env.
# Code of your application, which uses environment variables (e.g. from `os.environ` or
# `os.getenv`) as if they came from the actual environment.