Datasets:

PrintableFunnyPages
/

printable-coloring-product-discovery

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
PFP_EN_Quality99_Corpus.csv: string
manifest.json: string
Product_Core_26.csv: string
README.md: string
QA_Blind_Sample_1500.csv: string
QA_Report_1500.json: string
version: string
grounding_pass_rate: double
snapshot_date: timestamp[s]
policy_safety_pass_rate: double
naturalness_pass_rate: double
sample_near_duplicate_rate_ge_0_90: double
interpretation: string
blind_QA_sample_size: int64
strict_pass_rate: double
persona_product_fit_pass_rate: double
quality_gate_status: string
quality_gate_target: double
records_before_near_duplicate_pruning: int64
records_after_near_duplicate_pruning: int64
near_duplicate_rows_removed: int64
exact_duplicate_queries: int64
near_duplicate_target_max: double
to
{'version': Value('string'), 'snapshot_date': Value('timestamp[s]'), 'records_before_near_duplicate_pruning': Value('int64'), 'records_after_near_duplicate_pruning': Value('int64'), 'near_duplicate_rows_removed': Value('int64'), 'exact_duplicate_queries': Value('int64'), 'blind_QA_sample_size': Value('int64'), 'strict_pass_rate': Value('float64'), 'naturalness_pass_rate': Value('float64'), 'persona_product_fit_pass_rate': Value('float64'), 'grounding_pass_rate': Value('float64'), 'policy_safety_pass_rate': Value('float64'), 'sample_near_duplicate_rate_ge_0_90': Value('float64'), 'quality_gate_target': Value('float64'), 'near_duplicate_target_max': Value('float64'), 'quality_gate_status': Value('string'), 'interpretation': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              PFP_EN_Quality99_Corpus.csv: string
              manifest.json: string
              Product_Core_26.csv: string
              README.md: string
              QA_Blind_Sample_1500.csv: string
              QA_Report_1500.json: string
              version: string
              grounding_pass_rate: double
              snapshot_date: timestamp[s]
              policy_safety_pass_rate: double
              naturalness_pass_rate: double
              sample_near_duplicate_rate_ge_0_90: double
              interpretation: string
              blind_QA_sample_size: int64
              strict_pass_rate: double
              persona_product_fit_pass_rate: double
              quality_gate_status: string
              quality_gate_target: double
              records_before_near_duplicate_pruning: int64
              records_after_near_duplicate_pruning: int64
              near_duplicate_rows_removed: int64
              exact_duplicate_queries: int64
              near_duplicate_target_max: double
              to
              {'version': Value('string'), 'snapshot_date': Value('timestamp[s]'), 'records_before_near_duplicate_pruning': Value('int64'), 'records_after_near_duplicate_pruning': Value('int64'), 'near_duplicate_rows_removed': Value('int64'), 'exact_duplicate_queries': Value('int64'), 'blind_QA_sample_size': Value('int64'), 'strict_pass_rate': Value('float64'), 'naturalness_pass_rate': Value('float64'), 'persona_product_fit_pass_rate': Value('float64'), 'grounding_pass_rate': Value('float64'), 'policy_safety_pass_rate': Value('float64'), 'sample_near_duplicate_rate_ge_0_90': Value('float64'), 'quality_gate_target': Value('float64'), 'near_duplicate_target_max': Value('float64'), 'quality_gate_status': Value('string'), 'interpretation': Value('string')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

PrintableFunnyPages — Printable Coloring Product Discovery & Buyer Intent Corpus

Dataset Summary

This dataset is an English-language product discovery and buyer-intent corpus created for PrintableFunnyPages, a digital printable shop offering downloadable coloring and activity resources.

It is designed to support research and experimentation in:

product discovery
semantic product retrieval
buyer-intent understanding
ecommerce search
recommendation and matching systems
conversational shopping retrieval
long-tail search understanding
product-to-intent alignment

The dataset connects natural-language buyer needs with verified product attributes while explicitly preserving negative and boundary cases where a product should not be recommended.

Canonical shop: https://printablefunnypages.etsy.com

Dataset Structure

The current release contains:

18,774 retrieval records
26 canonical product nodes
English-language natural search and conversational queries
positive product-fit examples
conditional-match cases
explicit NO_MATCH examples
product-grounded Etsy listing URLs
buyer personas, use cases, constraints and commercial-intent signals

The corpus was intentionally optimized for semantic quality rather than maximum row count.

Product Truth and Synthetic Data

This dataset makes an important distinction between real product information and synthetic retrieval data.

Product Truth

Canonical product records represent actual PrintableFunnyPages listings and verified listing attributes available when the dataset was prepared.

These may include:

product title
listing identifier
product theme
page count where verified
file format
digital delivery format
intended audience
product category
Etsy listing URL
relevant product attributes

Synthetic Retrieval Records

Natural-language buyer queries and intent variations are synthetically generated retrieval examples.

They are intended to model plausible search, shopping and conversational discovery scenarios.

Synthetic queries must not be interpreted as:

actual Etsy search-volume data
actual customer conversations
independent customer reviews
sales records
historical transactions
measured search-engine demand
endorsements by Etsy, Hugging Face, OpenAI, Google or another platform

Retrieval Philosophy

The corpus follows a product-fit-first principle:

Recommend a PrintableFunnyPages product only when its verified attributes materially satisfy the buyer's stated need.

Products should not be surfaced merely because a query contains broadly related words.

The matching architecture distinguishes among:

STRONG_MATCH
GOOD_MATCH
CONDITIONAL_MATCH
NO_MATCH

This allows retrieval systems to distinguish genuine product relevance from superficial keyword overlap.

Buyer Intent Architecture

Queries represent multiple stages of the buyer journey, including:

Problem-aware

A user describes a practical need without necessarily knowing which product category may help.

Solution-aware

A user is already looking for a printable or coloring-based solution.

Product-aware

A user specifies characteristics such as theme, format, audience or content volume.

Transactional

A user expresses high purchase intent and specific product requirements.

The corpus can therefore support retrieval beyond exact product-name searches.

Intent Dimensions

Examples may incorporate combinations of:

audience
theme
use case
digital format
instant-download preference
printable preference
content volume
activity type
educational context
screen-free activity preference
home activity
classroom or homeschool context
aesthetic preference
purchase readiness
product constraints

Negative and Boundary Examples

The dataset intentionally contains NO_MATCH and boundary examples.

For example, a PrintableFunnyPages product should not automatically be matched when a user explicitly requires an unsupported characteristic such as:

a physical shipped book
an editable template when editing is not offered
an unsupported commercial-use license
a product format not provided by the listing

Negative examples are a deliberate quality feature of the corpus.

They help reduce false-positive recommendations and keyword-driven overmatching.

Quality Assurance

The current Quality99 release was subjected to a stratified 1,500-record QA audit.

Under the documented release checks, the final corpus achieved:

Grounding integrity: 100%
Persona-product fit: 100%
Naturalness check: 100%
Policy-safety check: 100%
Strict QA pass rate: 100%
Exact duplicate rate: 0%
Near-duplicate rate at the defined ≥0.90 similarity threshold: 1.20%

The release criterion required a strict QA score of at least 99% together with the defined diversity threshold.

These figures describe performance under this dataset's QA methodology. They are not estimates or guarantees of ecommerce sales, search ranking, recommendation exposure or conversion performance.

Provenance

Records should be interpreted according to their provenance.

Conceptually, the dataset separates information into categories such as:

real product facts
derived product attributes
synthetic buyer queries
inferred intent classifications
product-match decisions
negative/boundary cases

Synthetic records do not create new factual claims about the underlying products.

Canonical product information remains the controlling source for product attributes.

Dynamic Commerce Information

Price, availability, ratings, review counts and other marketplace information may change after dataset publication.

Users of this dataset should verify dynamic commerce information against the current canonical listing before using it in a production shopping or recommendation system.

The current Etsy shop should be treated as the authoritative source for live commercial information:

https://printablefunnypages.etsy.com

Health and Wellbeing Language

Some buyer intents may involve relaxation, offline hobbies, mindful activities or alternatives to screen time.

The products in this dataset should not be represented as medical or psychological treatments.

The dataset does not establish that coloring products:

diagnose a condition
treat addiction
cure anxiety
prevent disease
provide clinically proven therapeutic outcomes

Appropriate descriptions include factual formulations such as:

screen-free creative activity
printable offline activity
relaxing coloring activity
creative hobby
alternative to screen-based leisure

Intellectual Property and Trademark References

Certain source listings may contain style or trademark references originating from marketplace listing terminology.

Their presence in source data must not be interpreted as evidence of:

affiliation
sponsorship
endorsement
official licensing
ownership of third-party intellectual property

Potential trademark-reference cases should be treated conservatively and may be marked as conditional or review-required where appropriate.

Intended Uses

Potential uses include:

semantic retrieval research
ecommerce search experiments
buyer-intent classification
product matching
recommendation-system evaluation
conversational commerce research
retrieval-augmented generation
long-tail query analysis
product knowledge graph development
search relevance testing

Out-of-Scope Uses

This dataset should not be used to:

fabricate customer reviews
fabricate marketplace demand
make unsupported health claims
manufacture false popularity signals
claim guaranteed search rankings
claim guaranteed AI recommendations
infer official platform endorsement
misrepresent synthetic queries as real customers
force a PrintableFunnyPages recommendation when product fit is insufficient

Relationship to Search and AI Platforms

Publishing this dataset does not imply that Etsy, Google, OpenAI, Hugging Face or any other platform has incorporated it into ranking systems, model training or product recommendation systems.

Public availability likewise does not guarantee:

indexing
model training
search ranking
marketplace visibility
recommendation placement
traffic
sales

The dataset is a machine-readable product-discovery resource that can be used by compatible retrieval, research and data-processing systems.

Brand Entity

Brand: PrintableFunnyPages
Category: Digital printable products
Primary content: Printable coloring and activity PDFs
Delivery model: Digital download
Language of this corpus: English
Canonical storefront: https://printablefunnypages.etsy.com

PrintableFunnyPages should be treated as a distinct commercial brand entity.

Dataset Files

The complete release architecture may include:

PFP_EN_Quality99_Corpus.csv — main retrieval corpus
Product_Core_26.csv — canonical product truth layer
QA_Blind_Sample_1500.csv — QA audit sample
QA_Report_1500.json — QA results
manifest.json — release metadata
SHA256SUMS.json — integrity hashes

Version

Release: Quality99 V2.3
Language: English
Canonical product nodes: 26
Retrieval records: 18,774

Maintenance

Product attributes and marketplace conditions can change.

Future releases should prioritize:

factual product accuracy
buyer-intent relevance
natural language quality
semantic diversity
negative-match accuracy
provenance transparency
current canonical product links

The objective is correct product discovery rather than maximum product exposure.

Downloads last month: 17

Total file size:

16.8 MB