Text Realism & Enrichment

Misata generates domain-authentic, context-conditioned text across more than 15 real-world industry domains. It does not emit Latin placeholder gibberish ("Lorem Ipsum"), robotic single-sentence templates, or repetitive filler.

Every textual generator uses combinatorial grammar rules, seeded domain vocabulary pools, and cross-column context (e.g. sentiment aligned with star ratings, product descriptions conditioned on item categories, and clinical progress notes matching patient domains).

⚡ Quick Drop-in: misata.enrich_text()#

Have an existing pandas DataFrame, Series, or list of values with blank or boring text columns? Enrich them in one line with zero boilerplate:

import misata
import pandas as pd

# 1. Enrich an entire DataFrame with automatic semantic detection
df = pd.read_csv("support_cases.csv")
df_enriched = misata.enrich_text(df, seed=42)

# 2. Enrich a specific Series or column with an explicit text type
df["closing_notes"] = misata.enrich_text(
    df["closing_notes"],
    text_type="resolution_notes",
    seed=42,
)

# 3. Enrich a Python list or array
memos = misata.enrich_text([""] * 50, text_type="transaction_memo", seed=42)

Supported text_type Arguments#

Domain Showcase & Sample Outputs#

1. B2B SaaS & IT Helpdesk#

# Schema definition
columns = {
    "tickets": [
        Column(name="subject", type="text", distribution_params={"text_type": "ticket_subject"}),
        Column(name="description", type="text", distribution_params={"text_type": "support_ticket"}),
        Column(name="resolution_notes", type="text", distribution_params={"text_type": "resolution_notes"}),
    ]
}

Generated Samples:

• Subject: "Webhook 504 gateway timeout under load"
• Description: "The dashboard isn't loading any data since this morning. Several team members reported blank screens across both Chrome and Firefox."
• Resolution: "Root cause identified as database connection pool saturation. Deployed query index optimization and verified response latency normalized."

2. FinTech, Banking & Payments#

• Transaction Memo: "SWIFT WIRE REMITTANCE - INV #84920 REF:77391"
• Audit Reason: "Quarterly compliance review sign-off completed by chief risk officer"
• Account Statement Descriptor: "MERCHANT SETTLEMENT - STRIPE PAYMENTS BATCH #2910"

3. Healthcare & Clinical Informatics#

• Chief Complaint: "Persistent bilateral knee pain exacerbated by weight-bearing for 2 weeks"
• Discharge Summary: "Keep surgical incision clean and dry for 48 hours. Take prescribed antibiotics with food. Call clinic immediately if fever exceeds 101F."
• Progress Note (SOAP): "Patient presents for follow-up of hypertension. Vital signs stable. Medication well-tolerated. Continued current dosing with repeat lab work ordered."

4. E-Commerce & Retail#

Category-conditioned product descriptions automatically inspect the table's category column (or explicit parameter) to produce specs matching the product vertical:

# Electronics category:
# "Engineered with high-grade aluminum and precision components. Features advanced Bluetooth 5.3 connectivity with active noise cancellation and 30-hour battery life. Includes braided charging cable and hardshell travel case."

# Apparel category:
# "Crafted from breathable organic cotton blend with reinforced double-needle stitching. Designed with a modern tailored silhouette suitable for all-day comfort. Machine washable and pre-shrunk to retain fit."

• Return Reason: "Ordered wrong size or variant"
• Delivery Instructions: "Leave inside enclosed porch if raining, otherwise place on front step"

5. Sentiment-Calibrated Customer Reviews#

When a dataset includes both a rating column (1 to 5) and a review column, Misata couples the text generator to the rating value:

• 1 Star: "Extremely frustrated. Arrived damaged and customer service has not answered my emails for three days."
• 3 Stars: "Average quality for the price. Works as advertised but setup instructions were confusing."
• 5 Stars: "Exceptional build quality and exceeded all my expectations. Will definitely purchase again."

Using in Schemas#

Dict Schema#

schema = {
    "support_tickets": {
        "__rows__": 500,
        "id": {"type": "integer", "primary_key": True},
        "subject": {"type": "string", "text_type": "ticket_subject"},
        "resolution": {"type": "string", "text_type": "resolution_notes"},
        "error_trace": {"type": "string", "text_type": "error_message"},
        "churn_reason": {"type": "string", "text_type": "churn_reason"},
    }
}
tables = misata.generate_from_schema(misata.from_dict_schema(schema, seed=42))

YAML Schema#

tables:
  incidents:
    rows: 1000
    columns:
      incident_id: { type: int, unique: true }
      error_message: { type: text, text_type: error_message }
      resolution_notes: { type: text, text_type: resolution_notes }
      audit_reason: { type: text, text_type: audit_reason }