NLPTemplateEngine WL paclet

In brief

Wolfram Language (aka Mathematica) paclet for code generation by filling in templates using NLP techniques.

The paclet has a template fill in function Concretize and a related classifier of natural language specifications. A user can provide code templates to be filled in using CSV files.

The pre-loaded templates are used to produce Machine Learning (ML) computational workflow specifications for:

• Classification
• Latent Semantic Analysis
• Quantile Regression
• Recommendations
• Random tabular data generation

The implementations are in Python, R, and WL. Here is example of Python recommender creation spec:

---

Usage examples

Classification (WL)

Concretize["Quantile regression with the dataset dfTemp; use 20 knots and the probabilities 0.1 and 0.9."]

Hold[qrObj = 
   QRMonUnit[dfTemp]⟹
     QRMonEchoDataSummary[]⟹
     QRMonQuantileRegression[20, {0.1, 0.9}, InterpolationOrder -> 3]⟹
     QRMonPlot["DateListPlot" -> False, PlotTheme -> "Detailed"]⟹
     QRMonErrorPlots["RelativeErrors" -> False, "DateListPlot" -> False, PlotTheme -> "Detailed"];
]

Quantile Regression (WL)

Concretize["Quantile regression with the dataset dfTemp; use 20 knots and the probabilities 0.1 and 0.9."]

Hold[
 qrObj = QRMonUnit[dfTmep]⟹
     QRMonEchoDataSummary[]⟹
     QRMonQuantileRegression[20, {0.1, 0.9}, InterpolationOrder -> 3]⟹
     QRMonPlot["DateListPlot" -> False, PlotTheme -> "Detailed"]⟹
     QRMonErrorPlots["RelativeErrors" -> False, "DateListPlot" -> False, PlotTheme -> "Detailed"];
]

Recommendations (Python)

Concretize["Make a recommender over the data dfMovies; Give recommendations for the profile actor:Willis and year:1995.", 
 "TargetLanguage" -> "Python"]

smrObj = (SparseMatrixRecommender()
.create_from_wide_form(data = dfMovies, item_column_name="id", columns=None, add_tag_types_to_column_names=True, tag_value_separator=":")
.apply_term_weight_functions("IDF", "None", "Cosine")
.recommend_by_profile(profile=["Willis", "year"], nrecs=12)
.join_across(data=dfMovies, on="id")
.echo_value())

---

Workflow

TBD...

---

Targets

TBF...

Classification

Latent Semantic Analysis

Quantile Regression

Recommendations

Random tabular dataset

• Python,
• R,
• Raku
• WL, RandomTabularDataset

---

Bring your own templates

0. Load the paclet NLP-Template-Engine.

Needs["AntonAntonov`NLPTemplateEngine`"]

1. Get the "training" templates data (from CSV file you have created or changed) for a new workflow ("SendMail"):

dsSendMailTemplateEngineData = ResourceFunction["ImportCSVToDataset"][
  "https://raw.githubusercontent.com/antononcube/NLP-Template-Engine/main/TemplateData/dsQASParameters-SendMail.csv"];
Dimensions[dsSendMailTemplateEngineData]

(* {43, 5} *)

2. Add the ingested data for the new workflow (from the CSV file) into the NLP-Template-Engine:

NLPTemplateEngineAddData[dsSendMailTemplateEngineData] // Keys

(* {"Questions", "Templates", "Defaults", "Shortcuts"} *)

3. Parse natural language specification with the newly ingested and onboarded workflow ("SendMail"):

Concretize["SendMail", "Send email to joedoe@gmail.com with content RandomReal[343], and the subject this is a random real call.", PerformanceGoal -> "Speed"]

(* Hold[
 SendMail[
  Association["To" -> {"joedoe@gmail.com"}, 
   "Subject" -> "a random real call", "Body" -> RandomReal, 
   "AttachedFiles" -> None]]] *)

4. Experiment with running the generated code!

---

References

Repositories

[AAr1] Anton Antonov NLP Template Engine, (2021), GitHub/antononcube.

Videos

[AAv1] Anton Antonov "NLP Template Engine, Part 1", (2021), YouTube/AntonAntonov.

[AAv2] Anton Antonov, Natural Language Processing Template Engine, (2022), Wolfram Technology Conference 2022 presentation. YouTube/WolframResearch.