LAiSER Examples

This file contains copy-paste examples for the current refactored API.

1. Install From PyPI

pip install laiser

GPU-enabled install:

pip install "laiser[gpu]"

2. Skills Only From a Job Description

import os
import pandas as pd

from laiser.skill_extractor_refactored import SkillExtractorRefactored

data = pd.DataFrame(
    [
        {
            "Research ID": "job-001",
            "description": "Design scalable Python services and build analytics dashboards.",
        }
    ]
)

extractor = SkillExtractorRefactored(
    model_id="gemini",
    api_key=os.getenv("GEMINI_API_KEY") or os.getenv("GOOGLE_API_KEY"),
    use_gpu=False,
)

results = extractor.extract_concepts(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    input_type="job_desc",
    concepts=["skills"],
    top_k=5,
)

print(results)

3. Skills, Knowledge, and Tasks From a Job Description

results = extractor.extract_concepts(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    input_type="job_desc",
    concepts=["skills", "knowledge", "tasks"],
    top_k=5,
    allowed_sources=["esco", "onet"],
)

4. Course Syllabus Input

syllabus_df = pd.DataFrame(
    [
        {
            "Research ID": "course-001",
            "description": "Foundations of machine learning and data analysis.",
            "learning_outcomes": "Train models, interpret metrics, and communicate results.",
        }
    ]
)

results = extractor.extract_concepts(
    data=syllabus_df,
    id_column="Research ID",
    text_columns=["description", "learning_outcomes"],
    input_type="course_syllabi",
    concepts=["skills"],
)

Accepted syllabus aliases:

• syllabus
• course_syllabus
• course_syllabi

5. Restrict Taxonomy Sources

results = extractor.extract_concepts(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    input_type="job_desc",
    concepts=["skills"],
    allowed_sources=["esco"],
)

6. Write Output to CSV

results = extractor.extract_concepts(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    concepts=["skills", "knowledge", "tasks"],
    output_csv_path="alignment_results.csv",
)

7. Return Graph Edges

graph = extractor.extract_concepts(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    concepts=["skills", "knowledge", "tasks"],
    return_edges=True,
)

print(graph["nodes"].head())
print(graph["edges"].head())

8. Backward-Compatible Skills Wrapper

results = extractor.extract_and_align(
    data=data,
    id_column="Research ID",
    text_columns=["description"],
    input_type="job_desc",
)

9. Common Flags

Constructor:

• model_id
• hf_token
• api_key
• use_gpu
• backend

Extraction:

• input_type
• top_k
• similarity_threshold
• similarity_thresholds
• warnings
• allowed_sources
• concepts
• return_edges
• timing
• output_csv_path