Proto commits in chrislusf/gleam

These 55 commits are when the Protocol Buffers files have changed:

2018-08-15

format changes

The documentation is generated from this commit.

add step description

2018-07-13

set Driver.StopTime when driver process is killed before job finishes If the driver process is killed before the job finishes, the `Driver.StopTime` will be unassigned. The `Duration` field on the UI will keep growing. Signed-off-by: Shijiang Wei <mountkin@gmail.com>

2017-11-03

add Union Action for Datasets

2017-09-02

add cleaning up after each flow execution fix https://github.com/chrislusf/gleam/issues/58

2017-08-07

mapper reducer send stats to gleam executor

renaming

2017-08-06

adding flow name, improve master ui

2017-08-04

add name to each step

2017-08-02

pure go mapper/reducer now reports stats to agent

2017-07-30

temp commit before ready

2017-06-11

add new cassandra plugin, removing old adpter system

2017-05-14

move time to nanoseconds

2017-05-13

pass step id and task id to executor to track stats

2017-05-12

collect execution stat for each task

2017-05-09

[job status] adjust colors, add back start/end points, auto width, collect start/stop time

2017-05-08

[job status] displaying step groups with multiple sources

2017-05-03

WIP adding task group to job status

2017-05-01

master receives driver's flow status updates

2017-03-08

sql: support offset, fix logic for limit

2017-02-11

add distinct() function

2017-02-01

push files to agents directly

2017-01-02

better formatted progress when interrupted

driver collect execution status

2017-01-01

refactor pb message definition

2016-12-28

add pulse to avoid "connection reset by peer"

2016-12-27

remove LocalStatusRequest, StopRequest

2016-12-26

move execution and delete datasetshard to gRpc

2016-12-25

refactoring to merge the protobuf definitions

rename protobuf file names

adjust package location

2016-12-21

change to gRPC to work with master

2016-12-07

robust processing with OnDisk() option 1. retry on failed executions 2. only start execution when on disk inputs are ready 3. add errors when processing data

2016-12-02

add OnDisk() API

2016-11-29

hint for on disk data store

2016-11-25

profiling executors

2016-11-17

change all sorting to use *SortOption

2016-11-12

add cassandra and new plugin mechanism

2016-10-29

rename and move cmd package to msg

add ScatterPartitions instruction

2016-10-28

dynamic set dataset memory mode

2016-10-25

simplifying input output locations

2016-10-24

add distributed support for HashJoin()

add round robin execution

2016-10-22

added Top(n) and Take(n)

2016-10-19

add csv support, RoundRobin()

2016-10-14

add SortBy()

2016-10-13

add selective multiple key fields also add tf-idf example

2016-10-01

add counter for readers for efficient in memory channels

2016-09-28

wow, working in distributed mode!

2016-09-27

working in progress, adding distributed execution now it blocks when executing instructionset. the first line of data is read from the source netchan, but did not pass into the second instruction

2016-09-23

migrated distributed execution code from glow not working at all. just compiles fine.

2016-09-20

implement go executions on distributed executors

2016-09-18

add netchan

2016-09-17

moved distributed processing logic from glow change SaveTextTo() to Fprintf()