[BUG] tracing doesn't work when port is manually entered (ex: port=6007)

[kangeugine]

Describe the bug
Tracing doesn't work when port is manually entered through px.launch_app
The code below works fine when port is not manually entered (default=6006)

To Reproduce
Steps to reproduce the behavior:

import phoenix as px

from dotenv import load_dotenv, find_dotenv

from langchain.chat_models import ChatOpenAI
from langchain.prompts import ChatPromptTemplate
from langchain.chains import LLMChain
from phoenix.trace.langchain import OpenInferenceTracer, LangChainInstrumentor

load_dotenv(find_dotenv())
session = px.launch_app(port=6007) # HIGHLIGHT
tracer = OpenInferenceTracer()
LangChainInstrumentor(tracer).instrument()

llm = ChatOpenAI(temperature=0.9, model="gpt-3.5-turbo")
prompt = ChatPromptTemplate.from_template(
    "What is the best name to describe \
    a company that makes {product}?"
)
chain = LLMChain(llm=llm, prompt=prompt)
product = "Queen Size Sheet Set"
chain.run(product)

session.url

after running the above code, go to the url provided by session.url

Expected behavior
I would expect trace to exist for the query, prompt, and output

Screenshots
If applicable, add screenshots to help explain your problem.

Environment (please complete the following information):

• OS: Apple M1 (macOS Ventura 13.5.1)
• Browser: Chrome
• pheonix version: 2.4.1
• langchain version: 0.1.0

[RogerHYang]

@kangeugine Can you try setting it via environment variable instead?

import os

os.environ["PHOENIX_PORT"] = "6007"

[mikeldking]

soft deprecate params in launch_app and encourage everyone to use environment variables.