Skip to content

Latest commit

 

History

History
517 lines (331 loc) · 13.1 KB

CHANGES.rst

File metadata and controls

517 lines (331 loc) · 13.1 KB
Raw

Changelog and versioning

3.0.43

### Various fixes & improvements

  • Create a new printer for each query serialization (#198) by @snigdhas
  • feat(replay): add replays required timestamp col to ENTITY_TIME_COLUMNS (#197) by @aliu39
  • all-repos: update actions/upload-artifact to v4 (#191) by @joshuarli
  • expose extrapolate class to parent module (#196) by @enochtangg
  • feat(mql): Add extrapolate flag to MQL context (#195) by @enochtangg
  • fix: DeleteQuery, send query as a dict (#194) by @kylemumma
  • fix: add DeleteQuery to __init__ (#193) by @kylemumma
  • feat: new BaseQuery, DeleteQuery which will be used to make delete queries (#192) by @kylemumma
  • revert: sdfdsf (daf5d527) by @kylemumma
  • sdfdsf (bf73f803) by @kylemumma
  • revert: new DeleteQuery BaseQuery (0e5a47dc) by @kylemumma
  • new DeleteQuery BaseQuery (0fffebf4) by @kylemumma

2.0.35

### Various fixes & improvements

  • feat(metrics): add suffix wildcards & corresponding tests to filters in MQL grammar (#187) by @shellmayr
  • type tenant_ids as Mapping for variance (#188) by @asottile-sentry

2.0.34

### Various fixes & improvements

  • feat: Add support for storage queries.

2.0.33

### Various fixes & improvements

  • feat(ddm): Add lenient MRI regex (#184) by @iambriccardo

2.0.32

### Various fixes & improvements

  • feat: OrOptimizer (#183) by @kylemumma

2.0.31

  • Add support for generating correct unary expressions in MQL.

2.0.30

### Various fixes & improvements

  • feat(mql): Add support for unary operator (#179) by @iambriccardo

2.0.29

  • Fix filters grammar to avoid ambiguity

2.0.28

  • Fix a bug in how dependencies were defined to fix build bugs

2.0.27

  • Remove the entity field from the MQLContext
    • This also fixes a bug with using formulas in raw MQL
  • Update the CI to use 3.11 to stay compatible with Sentry
  • Remove references to tox, that is no longer used by CI

2.0.26

  • Support multiple entities in MQL Context
  • Support MQL strings as a query

2.0.25

  • Add support for or and and lowercase boolean operators in the MQL grammar.
  • Align MRI representation with backend, by allowing numbers in the metric name.
  • Add support for uppercase metric names.

2.0.24

  • Drop support for translating MetricsQuery to SnQL

2.0.23

  • Support empty filters in MQL

2.0.22

  • Support multiple parameters in curried arbitirary functions

2.0.21

  • Support curried arbitrary functions in MQL
  • Use tighter rules around aggregates, curried_aggregates, arbitrary_functions, and curried_arbitrary_functions in MQL

2.0.20

  • Allow arbitrary name for aggregate or curried aggreate in MQL
  • Fix bug in how the groupby in Formulas was being compared to their parameters

2.0.19

  • Fix escaping quotes in MQL strings

2.0.18

  • Support serializing Formula objects into MQL
  • Enable all the tests for deserializing MQL into Formulas
  • Support arbitrary functions in MQL

2.0.17

  • fix(ddm): Add support for . in the tag keys (#153) by @iambriccardo

2.0.16

  • Properly serialize curried aggregates in MQL

2.0.15

  • Fix a bug where quotes in MQL encoded strings were using single quotes instead of double quotes

2.0.14

  • Simplify the MQLContext and change it to serialize values other than strings

2.0.13

  • Expose a serialize_to_mql function on the Request object
    • This function will serialize MetricsQuery objects to MQL strings

2.0.12

  • Extend MQL grammar to support OR filters and parentheses

2.0.11

  • Add MQLContext to capture addition query information that cannot be expressed in MQL string
    • Add visitors to support MQLContext serialization
  • Support curried functions in the MQL grammar

2.0.10

  • Extend the MQL grammar to support unquoted tag values and the usage of spaces as delimiters in the filter string

2.0.9

  • Add a basic version of Formula SnQL translator
    • This translator is not meant to be permanent, it should be replaced with a MQL translator
  • Add MetricsQuery to MQL string encoder

2.0.8

  • Support the limit and offset field in the MetricsQuery
    • This will set the appropriate LIMIT and OFFSET clauses in the resulting SQL query,

    allowing for pagination

  • Align MQL filters grammar with Discover filters grammar.

2.0.7

  • Introduce MQL grammar and parser
  • Add Formula class to support formula queries.
  • Removed filters and groupby in MetricsQuery class

2.0.6

  • Make the granularity field optional in the Rollup, so that it can be automatically inferred by the API layer

2.0.5

  • Support the TOTALS clause in the Rollup of a MetricsQuery
  • Allow 10s granularity in MetricsQuery

2.0.4

  • Fix a bug where the groupby on the MetricsQuery was not being serialized

2.0.3

  • Bug fixes and changes to support using the MetricsQuery in production
  • Add granularity to the Rollup to delineate the interval (the results
bucketing) vs. the granularity of the underlying data
  • granularity is mandatory for now
  • This means adding back the restrictions of having interval or totals but not both
  • Use the interval to rollup the time automatically
    • This also means supporting arbitrary intervals
    • Automatically order timeseries by the rolled up time
  • Allow AliasedExpression in the groupby clause. This is to support resolving
    in Sentry. The result of the query should have the original tag key, not the resolved tag number. With AliasedExpression we can have Clickhouse do the mapping automatically.

2.0.2

  • Add validators and translators for the MetricsQuery and all its child objects
  • Rename some components to be consistent between Metric and Metrics
  • Move MetricsScope and Rollup to the timeseries module
    • Also stop treating them as Expressions, so they can be visited in a more nuanced way

2.0.1

### Various fixes & improvements

  • add additional setters to metric query child classes (#119) by @enochtangg
  • update contributing to reflect new release procedure (#118) by @evanh

2.0.0

  • Add a new syntax for querying timeseries metrics
    • Add a MetricsQuery class that can be attached to a Request
    • Add a Timeseries class that is used to capture a simple timeseries query
    • Add various other classes to support the new syntax

1.0.5

  • tenant_ids required field added previously is now optional to support gradual adoption

1.0.4

  • Add tenant_ids required field to Request object so additional information about the request can be passed from Sentry.
    • "tenants" include referrer, organization ID, etc.

1.0.3

  • Add the search issues time column to the legacy parser

1.0.2

  • Add the parent_api back to the Request so it can be passed from Sentry.

1.0.1

  • Modify column_name_re to allow for @ and / characters in column names.

1.0.0

  • Add a Request class that is the main entry point for Snuba.
    • The Request class contains the dataset being queried, the Query being sent, and any flags on the request
    • Flags have been removed from the Query entirely and are now set on the Request
  • Add an "app_id" flag to send the AppID to Snuba
  • Renamed snuba function to serialize

0.1.5

  • Modify ALIAS_RE to allow for @ char

0.1.4

  • Fix for a bug in 0.1.2 where parent_api was incorrectly validated

0.1.3

  • Modify ALIAS_RE to allow round brackets
  • Surround alias in AliasedExpression and CurriedFunction with backticks on query visitor

when translating a Query Object to a SNQL Query

0.1.2

  • Add support for team and feature tags for attribution in Snuba

0.1.1

  • Add support for lambdas and identifiers, which in turn enable higher order functions like arrayMap.

0.1.0

  • Move to Python 3.8 and drop support for Python 3.6. Sentry is now using 3.8 so this library can upgrade as well.
    • Use __future__.annotations where necessary
  • Update all dependencies to latest and fix subsequent linting errors
  • Add Data Model concept to Entities for extra validation
  • Create a type alias for sequences of conditions
  • Use sequences for LIMIT BY and ARRAY JOIN since Snuba now supports those operations over multiple columns

0.0.26

  • New release to fix dataclasses import issue with 3.8

0.0.25

  • Add a parent_api flag that is used to track the name of the calling API

0.0.24

  • Fix a bug in legacy converter that correctly handles infix conditions inside other functions

0.0.23

  • Add an AliasedExpression class that is used if the expression is in the select or groupby, which allows an alias of the results returned from Snuba. The alias is not used in any other clauses and is not available in the generated query in Snuba.
  • Fix ALIAS_RE to allow single letter aliases
  • Allow datetimes in legacy function strings (add : and - to allowed character list)

0.0.22

  • Allow square brackets in aliases

0.0.21

  • Try to convert wrapped conditions on tags to always use string comparisons.

0.0.20

  • Remove brittle, inconsistent and incomplete group by checks.

0.0.19

  • Fix escaping in queries. Move escaping from the legacy parser to the translator.

0.0.18

  • Add some more allowed characters to the function regex

0.0.17

  • Allow importing directly from snuba_sdk, e.g. from snuba_sdk import Column, Function
  • Fix bug where conditions on releases were being incorrectly parsed.

0.0.16

  • Fix bug with weirdly escaped slashes

0.0.15

Features:
  • Support embedding expressions in lists/tuples
  • Add a "legacy" flag that gets sent to Snuba for tracking
Fixes:
  • Fix for "+" in numbers
  • Don't strip more than the outer quotes on a string
  • Strip backticks out of strings from legacy queries

0.0.14

  • Add isort

0.0.13

  • Remove "transform" as an aggregate function
  • Fix for legacy queries that have raw string functions with nested aggregates
  • Stop conditions on tags[...] from being converted to dates if the rhs was a date string.
  • Some legacy queries use tuples for some of the fields. Convert them to lists where appropriate.

0.0.12

  • Small fix for legacy queries

0.0.11

  • Add arrayjoin support
  • Expand column regex to allow any tag

0.0.10

  • Handle sets in legacy queries

0.0.9

  • Fix a bug with aliases in legacy queries

0.0.8

  • Add support for dry run flag

0.0.7

  • Update the Legacy converter to produce Queries that will emulate the original query as closely as possible. This way we can track migration success by ensuring the SDK and legacy calls are both producing the same Clickhouse SQL.

0.0.6

  • Added documentation for the SDK, generated from Sphinx. The docs are located at https://getsentry.github.io/snuba-sdk/.
  • The SDK is now feature compatible with the legacy JSON Snuba API, so anything that can be done with the JSON can be done using this SDK.

0.0.5

  • Some small bug fixes uncovered after doing an integration test with Sentry and Snuba.

0.0.4

  • This package was originally developed for Python 3.8+, however Sentry (the main user) is still using Python 3.6.
  • The tox tests being used in Github Actions were not executing correctly, and so didn't report that this package was incompatible with 3.6.
  • The package was refactored to work with Python 3.6, and the tox tests were removed from Github Actions (but left in for easier local testing).

0.0.3

  • Add a query visitor for printing, validating and translating a Query object to a SnQL query
  • Allow arrays and tuples of scalars in Expressions
  • Add a function that translates from JSON snuba to SnQL SDK

0.0.2

  • It is now possible to create a functioning Query, with basic validation. Also CI and release tools have all been set up.

0.0.1

  • Created blank repo with basic bootstrapping

Versioning Policy

This project follows [semver](https://semver.org/), with three additions:

  • Semver says that major version 0 can include breaking changes at any time. Still, it is common practice to assume that only 0.x releases (minor versions) can contain breaking changes while 0.x.y releases (patch versions) are used for backwards-compatible changes (bugfixes and features). This project also follows that practice.
  • All undocumented APIs are considered internal. They are not part of this contract.
  • Certain features may be explicitly called out as "experimental" or "unstable" in the documentation. They come with their own versioning policy described in the documentation.

We recommend to pin your version requirements against 0.x.* or 0.x.y. Either one of the following is fine:

snuba-sdk>=0.10.0,<0.11.0
snuba-sdk==0.10.1

A major release N implies the previous release N-1 will no longer receive updates. We generally do not backport bugfixes to older versions unless they are security relevant. However, feel free to ask for backports of specific commits on the bugtracker.