knowledge/semantic/chunks
Source: knowledge/semantic/chunks.parquet. Source vintage: Package vintage 2026-10-01; source path partition not dated; inspect record and rule clocks separately. Grain: Packaged source record; inspect schema and linked world method before aggregation. Snapshot: 7a53af6a9beb58928d4a64dc92db09c9db65433f9909f1a10d16e302adaf86ba.
Read the method for this source
Explore this source with typed OData queries · Filtering, grouping and evidence method
Schema and filter instructions · Count this selection · First records
| node_id Int64 | document_node_id Int64 | doc_id Utf8 | ordinal Int64 | source_sha256 Utf8 | chunk_sha256 Utf8 | byte_start Int64 | byte_end Int64 | char_start Int64 | char_end Int64 | text Utf8 | extractor_version Utf8 | chunker_version Utf8 | __row_id Int64 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 10 Filter to this value Explore graph node | 10 Filter to this value | cms-cps-original-medicare-enrollment Filter to this value | 0 Filter to this value | 1dae5cf20e328f73835efc65f5f5a7e6af1220da50dd45cef67e30dc561f456d Filter to this value | a310f783eff5ec8a0a1be1221268abbabc19a1149d671ae9969e970a64ba842d Filter to this value | 0 Filter to this value | 691 Filter to this value | 0 Filter to this value | 691 Filter to this value | source: cms-cps-original-medicare-enrollment state: available table_state: raw_only gavel vintage: 2023-01-01 upstream files: MDCR%20ENROLL%20AB%209-14_CPS_02ENR_2023.zip why not shipped: this world does not carry it, and gavel holds no `data.parquet` for it either. gavel holds only the payload it fetched, under `_raw/`, and nothing was opened. Nobody converted the file. At vintage 2023-01-01 the source holds: MDCR%20ENROLL%20AB%209-14_CPS_02ENR_2023.zip. That is a conversion task in gavel and it is a different fact from a decision not to ship. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 10 |
| 11 Filter to this value Explore graph node | 11 Filter to this value | cms-hospital-compare Filter to this value | 0 Filter to this value | 703fc2ba7b53cd4431b7c66bee18d0eafabf0d1bfdcc8f3680cb123545c5fc25 Filter to this value | fcca008f6abb03e9ed466fb293d3f325b726d6215c0c9add07cc38fbd63550cf Filter to this value | 0 Filter to this value | 664 Filter to this value | 0 Filter to this value | 664 Filter to this value | source: cms-hospital-compare state: shipped table_state: converted gavel vintage: 2026-08-03 upstream files: data.parquet, cms_complications_ny.json named by: pipelines/coverage-intelligence/build_hospital_variance_extract.py names the object, pipelines/coverage-intelligence/probe_hospital_compare.py names the object gavel row count: 3240 note: Provider Data Catalog, complications and deaths 20 columns, including: facility_id, facility_name, address, citytown, state, zip_code, countyparish, telephone_number, measure_id, measure_name, compared_to_national, denominator, score, lower_estimate, higher_estimate, footnote, start_date, end_date, filename, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 11 |
| 12 Filter to this value Explore graph node | 12 Filter to this value | cms-hospital-general-information Filter to this value | 0 Filter to this value | 6c0e3c955bab558063540f1eb397f62678bea35fa20134b97078ce814dd489dc Filter to this value | dd507359a6e0708fee63ec091bd2dda77b7c0e8859bccb6cab9fbec54221946e Filter to this value | 0 Filter to this value | 1111 Filter to this value | 0 Filter to this value | 1111 Filter to this value | source: cms-hospital-general-information state: available table_state: converted gavel vintage: 2026-07-22 upstream files: data.parquet, Hospital_General_Information.csv gavel row count: 5419 note: [commons/open] Hospital General Information from the Provider Data Catalog: overall star rating and ownership per hospital CCN. Joins to the hospital standard-charge files by CCN. Fetched by `gavel_refresh commons`, resolver=pdc. why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 40 columns, including: Facility ID, Facility Name, Address, City/Town, State, ZIP Code, County/Parish, Telephone Number, Hospital Type, Hospital Ownership, Emergency Services, Meets criteria for birthing friendly designation, Hospital overall rating, Hospital overall rating footnote, MORT Group Measure Count, Count of Facility MORT Measures, Count of MORT Measures Better, Count of MORT Measures No Different and 22 more | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 12 |
| 13 Filter to this value Explore graph node | 13 Filter to this value | cms-ipps-msdrg-gmlos Filter to this value | 0 Filter to this value | 959d562c4bf320a408824e303b55eaf886a641f3ed1506fd70ced9d76b9f5fe8 Filter to this value | 460ed13420b68d01cd67793b61288bb920325b862d26e5d6b4253d4577ed05eb Filter to this value | 0 Filter to this value | 801 Filter to this value | 0 Filter to this value | 801 Filter to this value | source: cms-ipps-msdrg-gmlos state: available table_state: converted gavel vintage: 2027-01-01 upstream files: data.parquet gavel row count: 768 note: IPPS final rule Table 5: MS-DRG, post-acute and special-pay flags, MDC, type, title, relative weights, geometric and arithmetic mean LOS. The national expected LOS per MS-DRG why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 11 columns, including: MS-DRG, FY 2027 Final Post-Acute DRG, FY 2027 Final Special Pay DRG, MDC, TYPE, MS-DRG Title, Weights - Before Cap, Weights - 10% Cap Applied, Geometric mean LOS, Arithmetic mean LOS, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 13 |
| 14 Filter to this value Explore graph node | 14 Filter to this value | cms-ipps-table-5 Filter to this value | 0 Filter to this value | 8497b6a88aacd36860b954487d759bcf45a87655d935402761ee4d85ad518718 Filter to this value | be8063d0456034a7ccdaf1e02dd6551823b75254a7663e847162560b8dbb76d3 Filter to this value | 0 Filter to this value | 685 Filter to this value | 0 Filter to this value | 685 Filter to this value | source: cms-ipps-table-5 state: available table_state: raw_only gavel vintage: 2027-01-01 upstream files: fy2027-ipps-fr-table-5.zip, CMS-1849-F Table 5.xlsx why not shipped: this world does not carry it, and gavel holds no `data.parquet` for it either. gavel holds only the payload it fetched, under `_raw/`, and nothing was opened. Nobody converted the file. At vintage 2027-01-01 the source holds: fy2027-ipps-fr-table-5.zip, CMS-1849-F Table 5.xlsx. That is a conversion task in gavel and it is a different fact from a decision not to ship. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 14 |
| 15 Filter to this value Explore graph node | 15 Filter to this value | cms-ma-star-ratings Filter to this value | 0 Filter to this value | 25776faeeb25b60f273d4a549761473113a587fb561370b5ab6f3c38f1d7ebe8 Filter to this value | eede3f51c2f9eb83ae6fe59d39c7f362267703444f9458a66f2f0e3690fc34b4 Filter to this value | 0 Filter to this value | 1032 Filter to this value | 0 Filter to this value | 1032 Filter to this value | source: cms-ma-star-ratings state: available table_state: unpacked gavel vintage: 2026-01-01 upstream files: 2026-star-ratings-data-tables.zip, _extract.json, cai.csv, cai.parquet, cut-points.parquet, disaster.parquet, disenrollment-reasons.csv, disenrollment.parquet why not shipped: this world does not carry it, and gavel holds no `data.parquet` for it either. gavel opened the archive and holds its members under `_unpacked/`, and did not consolidate them into a `data.parquet`. The tables exist and are one join away, which is a smaller task than a conversion from scratch. At vintage 2026-01-01 the source holds: 2026-star-ratings-data-tables.zip, _extract.json, cai.csv, cai.parquet, cut-points.parquet, disaster.parquet, disenrollment-reasons.csv, disenrollment.parquet and 16 more object(s). That is a conversion task in gavel and it is a different fact from a decision not to ship. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 15 |
| 16 Filter to this value Explore graph node | 16 Filter to this value | cms-ma-star-ratings-cai Filter to this value | 0 Filter to this value | 5514219b97cd696a295477f6834ccd96b302b479cb33d265077393d9287e2ca1 Filter to this value | ff808752fbf735f22a6fa21805eaf2f72ef9f0eda65e1078b62caef3a9426b1c Filter to this value | 0 Filter to this value | 1103 Filter to this value | 0 Filter to this value | 1103 Filter to this value | source: cms-ma-star-ratings-cai state: available table_state: converted gavel vintage: 2026-01-01 upstream files: data.parquet gavel row count: 769 note: Categorical Adjustment Index. One row per contract, with the final adjustment category for Part C, Part D MA-PD, Part D PDP and Overall. This is the adjustment applied for a contract's share of dual-eligible and disabled enrollees, so it is the row that says a rating was adjusted and by which band why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 20 columns, including: star_year, contract_id, contract_name, organization_marketing_name, parent_organization, puerto_rico_only, part_c_fac_raw, part_c_fac, part_c_fac_not_reported, part_d_mapd_fac_raw, part_d_mapd_fac, part_d_mapd_fac_not_reported, part_d_pdp_fac_raw, part_d_pdp_fac, part_d_pdp_fac_not_reported, overall_fac_raw, overall_fac, overall_fac_not_reported, filename, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 16 |
| 17 Filter to this value Explore graph node | 17 Filter to this value | cms-ma-star-ratings-cut-points Filter to this value | 0 Filter to this value | 055cb975fa3db4447bed7228e58b14946fa46a4e91428c1a78b0f97b76cd7c35 Filter to this value | 2300245a0303e7f289ec36249f88b899af210aa84a30d3fa777ad5a8fbeb186f Filter to this value | 0 Filter to this value | 1034 Filter to this value | 0 Filter to this value | 1034 Filter to this value | source: cms-ma-star-ratings-cut-points state: available table_state: converted gavel vintage: 2026-01-01 upstream files: data.parquet gavel row count: 285 note: The thresholds that turn a rate into a star, Parts C and D in one table with a `part` column. One row per part per org type per star level per measure; 285 rows in the 2026 edition, being 5 x 33 for Part C and 10 x 12 for Part D, which additionally splits MA-PD from PDP. `threshold` is the PUBLISHED TEXT and is deliberately not parsed into bounds: the intervals come in four spellings and reverse direction for the measures where lower is better why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 12 columns, including: star_year, part, org_type, star_level, domain_code, domain_name, measure_code, measure_name, measurement_period, threshold, filename, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 17 |
| 18 Filter to this value Explore graph node | 18 Filter to this value | cms-ma-star-ratings-disaster Filter to this value | 0 Filter to this value | bb6ca95ef85676099ccb80dacf0587840e72125923f5326273a2e1e4b9e50bd5 Filter to this value | 14102dade3cd90c54eac7393245549e2da80aa9575656790d0611f406c3154ea Filter to this value | 0 Filter to this value | 841 Filter to this value | 0 Filter to this value | 841 Filter to this value | source: cms-ma-star-ratings-disaster state: available table_state: converted gavel vintage: 2026-01-01 upstream files: data.parquet gavel row count: 1538 note: Disaster percentages, one row per contract per DISASTER YEAR, which is not the edition year: the 2026 edition carries 2023 and 2024, so 769 x 2 = 1,538 rows. Long rather than wide because the set of years moves every edition. The value is a percentage written without a percent sign why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 8 columns, including: star_year, contract_id, disaster_year, disaster_pct_raw, disaster_pct, disaster_pct_not_reported, filename, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 18 |
| 19 Filter to this value Explore graph node | 19 Filter to this value | cms-ma-star-ratings-disenrollment Filter to this value | 0 Filter to this value | ab7bb15283dadad3d5d750448a5d55260971339bd0d875eb854eb0a57d6be82d Filter to this value | 585bfc5506ec7b5c11018904ecbf3d98ae60ec82f3c8c2a57ab8be87b3196ca6 Filter to this value | 0 Filter to this value | 869 Filter to this value | 0 Filter to this value | 869 Filter to this value | source: cms-ma-star-ratings-disenrollment state: available table_state: converted gavel vintage: 2026-01-01 upstream files: data.parquet gavel row count: 3845 note: Why members left, one row per contract per reason. 769 x 5 = 3,845 rows in the 2026 edition. `reason` is the published column label verbatim. The shares are percentages of disenrollees and 1,861 of the cells read `Not Available`, which is preserved rather than nulled why not shipped: this world does not carry it. The source is converted and readable in gavel and no tool here reads it. `available` means fetched and staged by gavel. It does not mean validated, does not mean converted, and does not mean read by anything here. 11 columns, including: star_year, contract_id, contract_name, organization_marketing_name, parent_organization, reason, share_raw, share, share_not_reported, filename, vintage | coverage-source-register/v1 Filter to this value | hammer-register-chunker-v3 Filter to this value | 19 |