BIBLIOTECA DE OPERADORES

El motor completo, un bloque a la vez

Cada operador es un bloque de construcción real y versionado. Busque, filtre por categoría y copie cualquier bloque directamente a su lienzo.

⌕

18 of 18 operatorsclick + Add to Canvas to copy

Retrieve

CORE UTILITY

Pulls data tables or stored models directly from repositories.

repository_entry (Path)

→ out (Data Table)

Read CSV / Excel

CORE UTILITY

Ingests external tabular structures.

file_encodingcolumn_separatorsfirst_row_names (Bool)

→ out (Data Table)

Read Database

HIGH RELEVANCE

Executes direct SQL queries against Snowflake, Postgres, etc.

connection_entryquery (SQL string)

→ out (Data Table)

Filter Examples

CORE UTILITY

Filters rows based on logical conditions.

condition_classparameter_string

→ ori (Original), fil (Filtered)

Select Attributes

CORE UTILITY

Retains or drops columns to reduce dimensions.

attribute_filter_type (subset, regex)

→ exa (Filtered Table)

Set Role

CORE UTILITY

Assigns special targets (e.g., label / prediction).

attribute_nametarget_role (label, id)

→ exa (Role-mapped Table)

Replace Missing Values

CORE UTILITY

Imputes empty cells natively.

default (average, minimum, zero, value)

→ exa (Imputed Table)

Join

CORE UTILITY

Horizontal table merges.

join_type (Inner, Left, Outer)key_attributes

→ join (Merged Table)

Random Forest

HIGH RELEVANCE

Ensemble model utilizing bagging decision trees.

number_of_treesmaximal_depthcriterion

→ mod (Trained Model)

Gradient Boosted Trees

HIGH RELEVANCE

Sequential loss-minimizing boosting trees.

number_of_treeslearning_rate

→ mod (Trained Model)

Apply Model

CORE UTILITY

Executes predictions on raw datasets using trained models.

None (auto-maps incoming columns)

→ lab (Predicted Table)

Cross Validation

HIGH RELEVANCE

Nested split-tester fold evaluator container.

number_of_folds (default: 10)sampling_type

→ mod, per (Metrics Table)

Performance (Classification)

CORE UTILITY

Scores categorical predictions.

accuracyprecisionrecallAUC (Bools)

→ per (Performance Metrics)

Execute Python

HIGH RELEVANCE

Runs embedded python scripting utilizing pandas DataFrames.

python_script (Code block)

→ out1, out2 (Pandas Outputs)

Segment Document

GENAI TREND

Splits massive unstructured documents into overlapping chunks.

segment_by (tokens, chars)chunk_sizechunk_overlap

→ doc (Segmented Chunks)

Embed & Index Documents

GENAI TREND

Generates numerical vector coordinates via selected models.

embedding_model_source (OpenAI, local)vector_dim

→ ind (Indexed Vector DB)

Retrieve Similar Docs

GENAI TREND

Semantic vector database querier.

top_ksimilarity_threshold

→ doc (Relevant Chunks)

Generate Text (LLM)

GENAI TREND

Passes dynamic context to LLM systems.

model_namesystem_prompttemperature

→ doc (LLM Generated Answer)