Abstract: A downhole recording device includes a housing centralizer, a housing, at least one radial connector connecting the housing to the housing centralizer, and an annular space between the housing centralizer and housing. The housing centralizer has a first end and a second end with a longitudinal opening through the housing centralizer from the first end to the second end. The housing is positioned radially within the longitudinal opening. The housing is configured to receive a downhole sensor. The annular space is located in the longitudinal opening of the housing centralizer and allows fluid communication from the first end of the housing centralizer to the second end of the housing centralizer.

Abstract: A downhole recording device includes a housing centralizer, a housing, at least one radial connector connecting the housing to the housing centralizer, and an annular space between the housing centralizer and housing. The housing centralizer has a first end and a second end with a longitudinal opening through the housing centralizer from the first end to the second end. The housing is positioned radially within the longitudinal opening. The housing is configured to receive a downhole sensor. The annular space is located in the longitudinal opening of the housing centralizer and allows fluid communication from the first end of the housing centralizer to the second end of the housing centralizer.

Abstract: Machine-learned models are selectively distributed to a plurality of computer servers according to conditions associated with the computer servers. A server receives travel information from a travel coordination system. The travel information describes a plurality of conditions. The server identifies a hierarchy of one or more parent-child relationships based on the plurality of conditions. The server trains machine-learned models using the plurality of conditions described by the travel information. The server selects machine-learned models for the plurality of conditions responsive to the identified hierarchy. The server distributes machine-learned models to the plurality of computer servers responsive to the identified hierarchy.

Abstract: Machine-learned models are selectively distributed to a plurality of computer servers according to conditions associated with the computer servers. A server receives travel information from a travel coordination system. The travel information describes a plurality of conditions. The server identifies a hierarchy of one or more parent-child relationships based on the plurality of conditions. The server trains machine-learned models using the plurality of conditions described by the travel information. The server selects machine-learned models for the plurality of conditions responsive to the identified hierarchy. The server distributes machine-learned models to the plurality of computer servers responsive to the identified hierarchy.

Abstract: A bit has a bit body. A pocket is formed in a surface of the bit body. The pocket has an opening at the surface and is defined by a rear wall with a sidewall extending between the rear wall and the surface. The pocket includes a stress relief feature at a front end of the sidewall to reduce stress concentrations in the bit body at or near the pocket, a stress relief feature at a rear end of the sidewall to reduce stress concentrations in the bit body at or near the pocket, or stress relief features at both front and rear ends of the sidewall.

Abstract: Porting between local, cluster, and cloud execution is aided by using the same source code for multiple data sources. A data source identifier recited in source code maps to data sources at different execution targets. Executable codes tailored to successive targets are produced using a resolution API without changing the developer's source code. An editable data source mapping is per-project, maintained in a file or project property, and placed in a distributable software package with executable code. Burdens on developers to determine execution targets in their source code and explicitly handle different execution locations in the source code are reduced. Source code can be freed of absolute path data source identifiers or code for detecting execution location. Source-to-source translation injects calls to create a folder, file, or cloud container. Data source resolution can be based on a relative path rooted at the data source identifier.

Abstract: Porting between local, cluster, and cloud execution is aided by using the same source code for multiple data sources. A data source identifier recited in source code maps to data sources at different execution targets. Executable codes tailored to successive targets are produced using a resolution API without changing the developer's source code. An editable data source mapping is per-project, maintained in a file or project property, and placed in a distributable software package with executable code. Burdens on developers to determine execution targets in their source code and explicitly handle different execution locations in the source code are reduced. Source code can be freed of absolute path data source identifiers or code for detecting execution location. Source-to-source translation injects calls to create a folder, file, or cloud container. Data source resolution can be based on a relative path rooted at the data source identifier.

Abstract: Computerized methods and systems for performing a search utilizing a spatial index are provided. Upon receiving a search query having a zoom level component (e.g., “Italian Restaurants in Seattle”), a search is initially conducted of the zoom level tags associated with the content items in the spatial index to determine if any of the content items include a zoom level tag indicative of the zoom level component of the search query (e.g., zoom level of city for Seattle). Subsequently, only those content items satisfying the zoom level search are grouped into a primary subset, whereby the primary subset may be filtered by selecting those returned content items with location-based tags that match a location-based component of the query (e.g., Seattle, Wash.). Next, the those content items satisfying the location-based filter are grouped into a location-based subset, which are filtered according to the parsed query components (e.g., category component Italian Restaurants).