Compare Java DSL to Python DSL

Consider porting the CSV DataSet so tests can be carried out using variables from a list and not only random/hardcoded values.

https://github.com/abstracta/jmeter-java-dsl/blob/master/jmeter-java-dsl/src/main/java/us/abstracta/jmeter/javadsl/JmeterDsl.java#L1359:1361

Having the name transferred to my pip account, (pypi/support#2260 (comment)), now I can use it for pymeter :)

Hi

It seems that pymeter uses jmeter-java-dsl 1.23.3 (as seen in pom.xml):

  <artifactId>jmeter-java-dsl</artifactId>
  <version>1.23.3</version>

As a developer I would like to declare custom variables.

Here is my current implementation

from typing import Optional

from pymeter.api import TestPlanChildElement

class Vars(TestPlanChildElement):
    def __init__(self, values: Optional[dict[str, str]]) -> None:
        self._vars_instance = TestPlanChildElement.jmeter_class.vars()
        if values:
            for k, v in values.items():
                self.set(k, v)
        super().__init__()

    def set(self, key: str, value: str):
        # TODO: perform type checking
        self._vars_instance = self.java_wrapped_element.set(key, value)
        return self

Usage Example:

variables = Vars({'api': config.get('api.endpoint')})

Add support to jmeter-influxdb2-listener-plugin (https://github.com/mderevyankoaqa/jmeter-influxdb2-listener-plugin) as it generates very useful metrics than the one already existing as it throws the requests raw data in the database so I think this would be a very useful feature to be added

This is more of a suggestion rather than an issue.

This is how I am defining my custom decorators so I can use them in my unit tests:

@inject
def http_get_performance_test(url: str,
                              threads=10,
                              duration=60,
                              html_reporter=Provide[Container.html_report]): # injected from dependency injection for a unified report
    def decorator(fun):
        def wrapper(*args, **kwargs):
            timer = UniformRandomTimer(1000, 2000)
            http_sampler = HttpSampler(
                fun.__name__,
                url,
                timer
            )
            tg = ThreadGroupWithRampUpAndHold(
                threads, 1, duration, http_sampler, name="Some Name"
            )
            tp = TestPlan(tg, html_reporter)
            fun(*args, **kwargs, test_plan=tp)

        return wrapper

    return decorator

When in use the unit tests look a lot cleaner:

class TestTestPlanClass(TestCase):
    @e2e.utils.decorators.http_get_performance_test(
        url="https://postman-echo.com/get?var=${__Random(0,10)}",
        threads=10,
        duration=30
    )
    def test_2(self, test_plan):
        stats = test_plan.run()
        self.assertLess(stats.sample_time_99_percentile_milliseconds, 2000)

I believe you can achieve a much nicer Developer Experience with decorators (you can also chain them)

Request:
Add support for the User Parameters preprocessor

Use case:
I have a JMeter test plan that uses User Parameters because the value needs to be different on each iteration. Due to the small size of the test, it performs better with User Parameters than with a csv data source.

Screenshot:

I would like to add websocket thread group support

Need to perform mvn dependency:copy-dependencies to download all jar dependencies and then to append them in the python package

Hi everyone!

I'm quite a newbie with Jmeter and pymeter. So, I'm trying to pass nested data via CSVDataSet.

What is a problem?
I have an application with an endpoint expecting list of string in body:

from typing import List

from pydantic import BaseModel
from fastapi import Request, FastAPI

app = FastAPI()

class Human(BaseModel):
    age: List[int]

# human: Human,
@app.post("/human/update")
async def update_human(human: Human, request: Request):
    a = await request.json()

And I have a pymeter TestPlan for such app:

local_data_schema = {
    "age": "${age}",
}

local_dcsv_data_set = CsvDataset("kek.csv")

http_sampler = (
    HttpSampler("echo_get_request", "http://0.0.0.0:8081/human/update")
    .post(local_data_schema, ContentType.APPLICATION_JSON)
)

thread_group = ThreadGroupSimple(
    1, 1, http_sampler, local_dcsv_data_set
)
test_plan = TestPlan(thread_group)
stats = test_plan.run()

with kek.csv like

,age
0, [1,2,3]

And all data that comes to andpoint is strings. As I see CSVDataSet uses , as delimeter, so even strings that comes to endoint are separated and value that comes first is '[1', so i can get the whole list. I've found the way to pass such nested data like python dict, but there is no option to send list of dicts, so I can't make my load test various.

I will be happy if you help me with that problem

Consider exposing children as a function in BaseThreadGroup (and other places where applicable).

Sometimes I don't want to pass the children in the init because I don't have them yet and would rather add them later.

Current workaround is to do something like this:

tg = ThreadGroupWithRampUpAndHold(10, 1, 60, name="Some Name")
tg._ramp_to_and_hold_instance.children(*[c.java_wrapped_element for c in children])

What I would like to see is:

tg = ThreadGroupWithRampUpAndHold(10, 1, 60, name="Some Name")
tg.children(children)

Hi

Nice, project, thanks for doing that.
I am curious, do you plan to add support for gRPC samples?
I will try to implement some custom samplers for POC.

I am attempting to access the response data from a sampler (generic in this example) via the JsonExtractor:

extractor = JsonExtractor("variable", "args.vendorname")
http_sampler = (HttpSampler("echo_get_request", "[https://postman-echo.com/post"](https://postman-echo.com/post%22), extractor)
                        .post("vendorname: ${vendorname}, description: ${description}, catalognumber: ${catalognumber}, size: 30, threshold: 0.0", ContentType.APPLICATION_JSON))

In the example in the docs I see that variable is accessed inside another sample like this:

dummy_sampler = DummySampler("dummy ${variable}", "hi dummy")

I tried to access that variable as a python obj, but nothing worked. Is there a way to read the value or access the response(s) from the HttpSampler directly?

i have been trying to send auth token and post data at the same time using this code

http_sampler = HttpSampler("sampler 1",url, timer,).header("Authorization", f"bearer {idToken}").post({"name":"abc"},ContentType.APPLICATION_JSON)

but it gives an error

AttributeError: 'us.abstracta.jmeter.javadsl.http.DslBaseHttpSample' object has no attribute 'post'

i tried to solve it myself and found that the value of self.java_wrapped_element when .header() is called is us.abstracta.jmeter.javadsl.http.DslHttpSampler but when .post() is called immediately after it, then the value changes to us.abstracta.jmeter.javadsl.http.DslBaseHttpSampler

but i am not good with classes so i couldn't find any solution for it