Sentiment Analyzer Operator

The Sentiment Analyzer LOP analyzes the sentiment of text input, providing scores and labels to indicate the emotional tone. It supports various analysis backends like Transformers, VADER, TextBlob, and Stanza, allowing you to choose the most suitable method for your needs. This operator can analyze full conversations, specific messages, or custom text, making it versatile for different use cases. The results are displayed in a table or summary format, offering insights into the sentiment trends of the analyzed text.

Requirements

• Python Packages:
  • transformers (if using Transformers backend)
  • vaderSentiment (if using VADER backend)
  • textblob (if using TextBlob backend)
  • stanza (if using Stanza backend) These can be installed via the ChatTD operator’s Python manager.
• ChatTD Operator: Required and must be configured (utilized for get_sentiment_backend).

Input/Output

Inputs

• Input Table (DAT): Table containing conversation data. Required columns: id, role, message.

Outputs

• Backend-Specific Table (DAT):
  • vader_table: VADER scores (compound, pos, neg, neu).
  • transformers_table: Transformers label and score.
  • textblob_table: TextBlob polarity and subjectivity.
  • stanza_table: Stanza sentiment score.
• Summary DAT: Table summarizing analysis results (average scores, etc.).

Parameters

Sentiment Page

Analyze Mode (Analyzemode) op('sentiment').par.Analyzemode Menu

Default:

full_conversation

Options:

full_conversation, last_message, specific_message, custom_text

Message Index (Messageindex) op('sentiment').par.Messageindex Integer

Default:

1

Custom Text (Customtext) op('sentiment').par.Customtext String

Default:

None

Start Analysis (Analyze) op('sentiment').par.Analyze Pulse

Default:

None

Clear Results (Clear) op('sentiment').par.Clear Pulse

Default:

None

Status (Status) op('sentiment').par.Status String

Default:

None

Display Mode (Display) op('sentiment').par.Display Menu

Default:

sentiment_table

Options:

sentiment_table, summary

Analysis Backend (Backend) op('sentiment').par.Backend Menu

Default:

transformers

Options:

transformers, vader, textblob, stanza

Model Name (Modelname) op('sentiment').par.Modelname StrMenu

Default:

distilbert-base-uncased-finetuned-sst-2-english

Language (Language) op('sentiment').par.Language Menu

Default:

auto

Options:

auto, en, es, fr, de, it

Table Update Mode (Updatemode) op('sentiment').par.Updatemode Menu

Default:

append

Options:

append, replace, batch, clear

About Page

ChatTD Operator (Chattd) op('sentiment').par.Chattd OP

Default:

None

Bypass (Bypass) op('sentiment').par.Bypass Toggle

Default:

false

Show Built-in Parameters (Showbuiltin) op('sentiment').par.Showbuiltin Toggle

Default:

false

Version (Version) op('sentiment').par.Version String

Default:

None

Last Updated (Lastupdated) op('sentiment').par.Lastupdated String

Default:

None

Creator (Creator) op('sentiment').par.Creator String

Default:

None

Website (Website) op('sentiment').par.Website String

Default:

None

Performance Considerations

• Backend Choice: Transformers models are generally more accurate but slower than VADER or TextBlob.
• Input Size: Analyzing large conversations/texts takes longer. Use last_message or specific_message modes if performance is critical.

Usage Examples

Analyzing a Full Conversation (VADER)

sentiment_op = op('sentiment1')
conv_dat = op('conversation_log')

sentiment_op.inputConnectors[0].connect(conv_dat)
sentiment_op.par.Analyzemode = 'full_conversation'
sentiment_op.par.Backend = 'vader'
sentiment_op.par.Analyze.pulse()

# Results in sentiment_op.op('vader_table')

Analyzing Custom Text (Transformers)

sentiment_op = op('sentiment1')

sentiment_op.par.Analyzemode = 'custom_text'
sentiment_op.par.Customtext = "This movie was absolutely fantastic! Highly recommended."
sentiment_op.par.Backend = 'transformers'
sentiment_op.par.Modelname = 'distilbert-base-uncased-finetuned-sst-2-english' # Default is good for general English
sentiment_op.par.Analyze.pulse()

# Results in sentiment_op.op('transformers_table')

Displaying Summary

sentiment_op = op('sentiment1')
# ... perform analysis first ...
sentiment_op.par.Display = 'summary'

# Summary results in sentiment_op.op('summary_dat')

Common Use Cases

• Analyzing customer feedback sentiment.
• Monitoring social media for public opinion.
• Tracking emotional trends in chat logs.
• Real-time sentiment display in chatbots.
• Flagging negative messages.