Skip to content

sentiment analysis on translation

Open In Colab Youtube

Multi-Module Pipeline: Sentiment Analysis on Translation

🇨🇴 Versión en español de este documento

This document details a multi-modular pipeline that takes in an a text file in a non-English language, translates it into English, and then performs sentiment analysis on each sentence of the translation.

Performing sentiment analysis on translations helps users understand the emotional tone of translated text, aiding in cross-cultural communication, market research, public opinion measurement, and customer feedback analysis across different languages, among other possibilities.

The document is divided into the following sections:

Pipeline Setup

To achieve what we've described above, let's set up a pipeline sequentially consisting of the following modules:

We do this by leveraging the create_pipeline method, as follows:

# create a pipeline as detailed above
pipeline = krixik.create_pipeline(name="multi_sentiment_analysis_on_translation", module_chain=["parser", "translate", "sentiment"])

Processing an Input File

A pipeline's valid input formats are determined by its first module—in this case, a parser module. Therefore, this pipeline only accepts textual document inputs.

Given that we're translating and then performing sentiment analysis, we'll start with a file in Spanish. Let's take a quick look at the file before proceeding (corrupted accent marks are a quirk of the print mechanism, but they do not thus flow into the AI model):

# examine contents of the input file
with open(data_dir + "input/spanish_review.txt", "r") as file:
    print(file.read())

Para los trabajos que estoy haciendo me resultó muy bueno. En una hora carga la batería y dura más de 3 horas de trabajo continuo. Un golazo contar con una segunda batería. Cómodo y con buen torque. Estoy conforme.

Since the input text is in Spanish, we'll use the (non-default) opus-mt-es-en model of the translate module to translate it into English.

We will use the default models for every other module in the pipeline, so they don't have to be specified in the modules argument of the process method.

# process the file through the pipeline, as described above
process_output = pipeline.process(
    local_file_path=data_dir + "input/spanish_review.txt",  # the initial local filepath where the input file is stored
    local_save_directory=data_dir + "output",  # the local directory that the output file will be saved to
    expire_time=60 * 30,  # process data will be deleted from the Krixik system in 30 minutes
    wait_for_process=True,  # wait for process to complete before returning IDE control to user
    verbose=False,  # do not display process update printouts upon running code
    modules={"module_2": {"model": "opus-mt-es-en"}},
)  # specify a non-default model for use in the second module

The output of this process is printed below. To learn more about each component of the output, review documentation for the process method.

Because the output of this particular module-model pair is a JSON file, the process output is provided in this object as well (this is only the case for JSON outputs). Moreover, the output file itself has been saved to the location noted in the process_output_files key. The file_id of the processed input is used as a filename prefix for the output file.

# nicely print the output of this process
print(json.dumps(process_output, indent=2))

{
  "status_code": 200,
  "pipeline": "multi_sentiment_analysis_on_translation",
  "request_id": "0719e2d3-24cc-4a12-9f84-d90d1a9e964e",
  "file_id": "ed5b8fa4-3f4b-4e10-b61d-05536db5e929",
  "message": "SUCCESS - output fetched for file_id ed5b8fa4-3f4b-4e10-b61d-05536db5e929.Output saved to location(s) listed in process_output_files.",
  "warnings": [],
  "process_output": [
    {
      "snippet": "For the jobs I'm doing I turned out very good.",
      "positive": 1.0,
      "negative": 0.0,
      "neutral": 0.0
    },
    {
      "snippet": "In one hour load the battery and last more than 3 hours of continuous work.",
      "positive": 0.008,
      "negative": 0.992,
      "neutral": 0.0
    },
    {
      "snippet": "A golasse to have a second batter.",
      "positive": 0.023,
      "negative": 0.977,
      "neutral": 0.0
    },
    {
      "snippet": "Cmodo and with good torque.",
      "positive": 1.0,
      "negative": 0.0,
      "neutral": 0.0
    },
    {
      "snippet": "I agree.",
      "positive": 1.0,
      "negative": 0.0,
      "neutral": 0.0
    }
  ],
  "process_output_files": [
    "../../../data/output/ed5b8fa4-3f4b-4e10-b61d-05536db5e929.json"
  ]
}

To confirm that everything went as it should have, let's load in the text file output from process_output_files:

# load in process output from file
with open(process_output["process_output_files"][0]) as f:
    print(json.dumps(json.load(f), indent=2))

[
  {
    "snippet": "For the jobs I'm doing I turned out very good.",
    "positive": 1.0,
    "negative": 0.0,
    "neutral": 0.0
  },
  {
    "snippet": "In one hour load the battery and last more than 3 hours of continuous work.",
    "positive": 0.008,
    "negative": 0.992,
    "neutral": 0.0
  },
  {
    "snippet": "A golasse to have a second batter.",
    "positive": 0.023,
    "negative": 0.977,
    "neutral": 0.0
  },
  {
    "snippet": "Cmodo and with good torque.",
    "positive": 1.0,
    "negative": 0.0,
    "neutral": 0.0
  },
  {
    "snippet": "I agree.",
    "positive": 1.0,
    "negative": 0.0,
    "neutral": 0.0
  }
]

You may note that, in the first returned snippet, the word "sillón" is missing its second vowel and is printed as "silln". This is a model issue: the translate model with which we processed the file may have trouble with accented characters and/or outright remove them. It's important that you familiarize yourself with the peculiarities of AI models you intend to leverage often.