Support Plots.jl plots (#42)

* Remove lightgraphs dependency and use Requires.jl  ( fix #39 )

* Add Plots support
* Add PyPlot support

* Replace all `log_image(s)` methods with a common one, that internally use all our preprocess machinery (should use it for everything else too)
* move LogGraph to Optional folder

* BugFix: log_***(step=nothing) gave errors

* Update travis to Test on latest julia
* Docs

Author: PhilipVinc

.travis.yml

@@ -4,7 +4,7 @@ os:
     - linux
 julia:
     - 1.0
-    - 1.1
+    - 1
     - nightly
 matrix:
     allow_failures:

Project.toml

@@ -8,8 +8,8 @@ CRC32c = "8bf52ea8-c179-5cab-976a-9e18b702a9bc"
 ColorTypes = "3da002f7-5984-5a60-b8a6-cbb66c0b333f"
 FileIO = "5789e2e9-d7fb-5bc7-8068-2c6fae9b9549"
 ImageCore = "a09fc81d-aa75-5fe9-8630-4744c3626534"
-LightGraphs = "093fc24a-ae57-5d10-9952-331d41423f4d"
 ProtoBuf = "3349acd9-ac6a-5e09-bcdb-63829b23a429"
+Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
 
 [compat]
@@ -20,10 +20,13 @@ julia = "0.7, 1"
 [extras]
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"
 ImageMagick = "6218d12a-5da1-5696-b52f-db25d2ecc6d1"
+LightGraphs = "093fc24a-ae57-5d10-9952-331d41423f4d"
 Logging = "56ddb016-857b-54e1-b83d-db4d58db5568"
+Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
+PyPlot="d330b81b-6aea-500a-939a-2ce795aea3ee"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 TestImages = "5e47fb64-e119-507b-a336-dd2b206d9990"
 WAV = "8149f6b0-98f6-5db9-b78f-408fbbb8ef88"
 
 [targets]
-test = ["Test", "Flux", "TestImages", "ImageMagick", "Logging", "WAV"]
+test = ["Test", "Flux", "TestImages", "ImageMagick", "Logging", "LightGraphs", "Plots", "PyPlot", "WAV"]

README.md

@@ -16,7 +16,8 @@ To install run the following command in the julia REPL:
 ] add TensorBoardLogger
 ```
 
-To also enable support for Images, also install `ImageMagick`
+If you need to log Images, you must also install `ImageMagick`
+(on MacOS, you will also need `QuartzImageIO`).
 ```
 ] add ImageMagick
 ```
@@ -34,6 +35,14 @@ logger in Julia:
     - You can set it to be the current logger in a scope with the function [`with_logger`](https://docs.julialang.org/en/v1/stdlib/Logging/index.html#Base.CoreLogging.with_logger)
     - You can combine it with other Loggers using [LoggingExtras.jl](https://github.com/oxinabox/LoggingExtras.jl), so that messages are logged to TensorBoard and to other backends at the same time.
 
+You can log to TensorBoard any type. Numeric types will be logged as scalar,
+arrays will be binned into histograms, images and audio will be logged as such,
+and we even support [Plots](https://github.com/JuliaPlots/Plots.jl) and
+[PyPlot](https://github.com/JuliaPlots/Plots.jl) figures!
+
+For details about how types are logged by default, or how to customize this behaviour for your custom types,
+refer to the documentation or the examples folder.
+
 ```julia
 using TensorBoardLogger, Logging, Random
 
@@ -59,6 +68,10 @@ with_logger(lg) do
 end
 ```
 
+## Integration with third party packages
+We also support native logging of the types defined by a few third-party packages, such as `Plots` and `PyPlot` plots.
+If there are other libraries that you think we should include in the list, please open an issue.
+
 ## Roadmap
 Support for Computational Graphs is currently planned and partly implemented.
 

docs/make.jl

@@ -8,9 +8,9 @@ makedocs(
     pages     = [
             "Home"          => "index.md",
             "Manual"        => Any[
-                "Basics"        => "index.md",
-                "Backends"    =>  "custom_behaviour.md",
-                "Extending"    =>  "extending_behaviour.md"
+                "Backends"     =>  "custom_behaviour.md",
+                "Extending"    =>  "extending_behaviour.md",
+                "Explicit Interface" => "explicit_interface.md"
             ],
             "Examples"      => Any[
                 "Optim.jl"  => "examples/optim.md"

docs/src/custom_behaviour.md

@@ -26,7 +26,6 @@ parameters. For example, to log a vector as text you can do the following:
 @info "mytag" TBText([1,2,3])
 ```
 
-
 ```@docs
 TBText
 TBHistogram

docs/src/explicit_interface.md

@@ -0,0 +1,35 @@
+# Explicit interface
+
+in addition to the standard logging interface, it is possible to log
+data to TensorBoard using the functions documented below.
+All the functions accept take as first argument a `TBLogger` object
+and as the second argument a `String` as the tag under which the
+data will be logged.
+
+# Scalar backend
+```@docs
+log_value
+```
+
+# Distributions backend
+```@docs
+log_histogram
+log_vector
+```
+
+# Text backend
+```@docs
+log_text
+```
+
+# Images backend
+```@docs
+log_image
+log_images
+```
+
+# Audio backend
+```@docs
+log_audio
+log_audios
+```

docs/src/index.md

@@ -15,8 +15,9 @@ To install this Julia package run the following command in the julia REPL:
 ```
 
 !!! note
-    To log images to the *Image* backend of TensorBoard you must also install
-    `ImageMagick`. You won't need to import it.
+    If you need to log Images, you must also install `ImageMagick`
+    (on MacOS, you will also need `QuartzImageIO`).
+    You only need to install those dependencies, and you will never need to load them.
     ```
     ] add ImageMagick
     ```
@@ -41,10 +42,17 @@ there already exhist a document at the given path.
 
 Once you have created a `TBLogger`, you can use it as you would use any other
 logger in Julia:
-    - You can set it to be your global logger with the function [`global_logger`](https://docs.julialang.org/en/v1/stdlib/Logging/index.html#Base.CoreLogging.global_logger)
-    - You can set it to be the current logger in a scope with the function [`with_logger`](https://docs.julialang.org/en/v1/stdlib/Logging/index.html#Base.CoreLogging.with_logger)
-    - You can combine it with other Loggers using [LoggingExtras.jl](https://github.com/oxinabox/LoggingExtras.jl), so that messages are logged to TensorBoard and to other backends at the same time.
+- You can set it to be your global logger with the function [`global_logger`](https://docs.julialang.org/en/v1/stdlib/Logging/index.html#Base.CoreLogging.global_logger)
+- You can set it to be the current logger in a scope with the function [`with_logger`](https://docs.julialang.org/en/v1/stdlib/Logging/index.html#Base.CoreLogging.with_logger)
+- You can combine it with other Loggers using [LoggingExtras.jl](https://github.com/oxinabox/LoggingExtras.jl), so that messages are logged to TensorBoard and to other backends at the same time.
 
+Every `TBLogger` has an internal counter to store the current `step`, which is initially set to `1`. All the data logged with the same `@log` call will be logged with the same step, and then
+it will increment the internal counter by 1.
+
+If you want to increase the counter by a different amount, or prevent it from increasing, you can log the additional message
+`log_step_increment=N`. The default behaviour corresponds to `N=1`. If you set `N=0` the internal counter will not be modified.
+
+See the example below:
 ```julia
 using TensorBoardLogger, Logging, Random
 
@@ -92,8 +100,13 @@ how to specify a desired backend refer to [Specifying a backend](@ref).
 If you want to define a new default behaviour for a custom type refer to section
 [Extending TensorBoardLogger](@ref).
 
+## Third-party packages
+We also support logging custom types from a the following third-party libraries:
+ - [Plots.jl](https://github.com/JuliaPlots/Plots.jl): the `Plots.Plot` type will be rendered to PNG at the resolution specified by the object and logged as an image
+ - [PyPlot.jl](https://github.com/JuliaPy/PyPlot.jl): the `PyPlot.Figure` type will be rendered to PNG at the resolution specified by the object and logged as an image
+
 ## Explicit logging
 
 In alternative, you can also log data to TensorBoard through it's functional interface,
 by calling the relevant method with a tag string and the data. For information
-on this interface refer to @ref()...
+on this interface refer to [Explicit interface]@ref()...

src/ImageFormat.jl

@@ -0,0 +1,16 @@
+# Note: we use the 100s place of the enum to represent the number of dimensions
+# And the 10s place to represent which dimension is the observation (N), 0 for no observation dimension
+@enum ImageFormat begin
+    L=100;
+    CL=200; LC; HW; WH; NL=210; LN=220;
+    HWC=300; WHC; CHW; CWH; NCL=310; NHW; NWH; NLC; CLN=330; LCN; HWN; WHN;
+    NHWC=410; NWHC; NCHW; NCWH; HWCN=440; WHCN; CHWN; CWHN;
+end
+
+expected_ndims(x::ImageFormat) = Int(x) ÷ 100 # get the 100's digit
+obs_dim(x::ImageFormat) = (Int(x) % 100) ÷ 10 # get the 10's digit
+const strip_obs = Dict(
+    NL=>L, LN=>L, NCL=>CL, CLN=>CL, NLC=>LC, LCN=>LC,
+    NHW=>HW, HWN=>HW, NWH=>WH, WHN=>WH,
+    NHWC=>HWC, NWHC=>WHC, NCHW=>CHW, NCWH=>CWH, HWCN=>HWC, WHCN=>WHC, CHWN=>CHW, CWHN=>CWH
+)

src/Loggers/LogAudio.jl

@@ -1,5 +1,5 @@
 """
-    log_audios(logger::TBLogger, name::AbstractString, samples::AbstractArray, samplerate::Real, step)
+    log_audios(logger::TBLogger, name::AbstractString, samples::AbstractArray, samplerate::Real; step=step(logger))
 
 Logs multiple audio clips at step `step`
 - samples:
@@ -12,7 +12,7 @@ function log_audios(logger::TBLogger, name::AbstractString, samplesArray::Abstra
     end
 end
 """
-    log_audio(logger::TBLogger, name::AbstractString, samples::AbstractArray, samplerate::Real, step)
+    log_audio(logger::TBLogger, name::AbstractString, samples::AbstractArray, samplerate::Real; step=step(logger))
 
 Logs an audio clip with name `name` at step `step`
 - samples: Array of samples N*C where N = number of samples and C = number of channels

src/Loggers/LogHistograms.jl

@@ -1,5 +1,5 @@
 """
-    log_histogram(logger, name, (bins,weights); step)
+    log_histogram(logger, name, (bins,weights); step=step(logger))
 
 Logs a histogram under the tag `name` on given `step`. The histogram must be
 passed as a tuple holding the `N+1` bin edges and the height of the `N` bins.
@@ -15,7 +15,7 @@ function log_histogram(logger::TBLogger, name::AbstractString, (bins,weights)::T
 end
 
 """
-    log_histogram(logger, name, data::Vector; step)
+    log_histogram(logger, name, data::Vector; step=step(logger))
 
 Bins the values found in `data` and logs them as an histogram under the tag
 `name`.
@@ -29,7 +29,7 @@ function log_histogram(logger::TBLogger, name::AbstractString, data::AbstractArr
 end
 
 """
-    log_vector(logger, name, data::Vector; step)
+    log_vector(logger, name, data::Vector; step=step(logger))
 
 Logs the vector found in `data` as an histogram under the name `name`.
 """