Update jars, fix link rot

.gitignore

@@ -1,6 +1,6 @@
-deps/tika-app-1.10.jar
+deps/*.jar
 deps/fop-2.0
-deps/fop-2.0-bin.tar.gz
+deps/*.tar.gz
 test/simple.pdf
 .DS_Store
 docs/build

deps/build.jl

@@ -1,15 +1,15 @@
 tdeps = dirname(@__FILE__)
-tika_jar = joinpath(tdeps, "tika-app-1.10.jar")
+tika_jar = joinpath(tdeps, "tika-app-1.17.jar")
 if !isfile(tika_jar)
-    info("  Downloading tika-app-1.10.jar from Maven Central")
-    download("http://search.maven.org/remotecontent?filepath=org/apache/tika/tika-app/1.10/tika-app-1.10.jar", tika_jar)
+    info("  Downloading tika-app-1.17.jar from Maven Central")
+    download("http://search.maven.org/remotecontent?filepath=org/apache/tika/tika-app/1.17/tika-app-1.17.jar", tika_jar)
 end
 
-fop_jar = joinpath(tdeps, "fop-2,0", "fop-2.0.jar")
-fop_gz = joinpath(tdeps, "fop-2.0-bin.tar.gz")
+fop_jar = joinpath(tdeps, "fop-2,2", "fop-2.2.jar")
+fop_gz = joinpath(tdeps, "fop-2.2-bin.tar.gz")
 
 if !isfile(fop_gz)
-    info("  Downloading fop-2.0 binary from Apache OSUOSL Mirror")
+    info("  Downloading fop-2.2 binary from Apache OSUOSL Mirror")
     download("http://apache.osuosl.org/xmlgraphics/fop/binaries/fop-2.0-bin.tar.gz", fop_gz)
 end
 if !isfile(fop_jar)

docs/src/guide/extract.md

@@ -5,7 +5,7 @@ Taro includes a few high level functions that extract data from various document
 ##Text extraction
 
 The [`Taro.extract`](@ref) method retrieves document metadata and the body text of a document,
-using [Apache Tika](https://tika.apache.org/). Formats [supported by Tika](https://tika.apache.org/1.13/formats.html)
+using [Apache Tika](https://tika.apache.org/). Formats [supported by Tika](https://tika.apache.org/1.17/formats.html)
 include MS Office and Open Office documents, as well as PDF files.
 
 The function return a Tuple of a Dict and String. The Dict contains name/value pairs of various metadata from the document, while the string contains the body text of the document.
@@ -21,12 +21,12 @@ text[1:53]
 ## Read Excel files into a DataFrame
 
 The [`Taro.readxl`](@ref) method reads a rectangular region from an excel sheet, and
-returns a [Dataframe](http://juliastats.github.io/DataFrames.jl/stable/man/getting_started/#The-DataFrame-Type-1).
+returns a [Dataframe](http://juliadata.github.io/DataFrames.jl/latest/man/getting_started.html#The-DataFrame-Type-1).
 This function takes as an input parameter the name and path of the Excel file to be read. A sheet name (or number) can optionally be supplied. If no sheet information is given, the first sheet (index 0) is read. Finally, this
 function is provided with the rectangular region from which data is extracted. This region is specified as an excel
 range.
 
-This function is similar to, and inspired by, the [readtable](http://juliastats.github.io/DataFrames.jl/stable/man/io/#DataFrames.readtable) function in DataFrames.
+This function is similar to, and inspired by, the [CSV.read/DataFrames.readtable](http://juliadata.github.io/CSV.jl/latest/#CSV.read) function in CSV/DataFrames.
 
 ```@repl
 using Taro # hide

docs/src/index.md

@@ -19,8 +19,8 @@ The [ExcelReaders.jl](https://github.com/davidanthoff/ExcelReaders.jl) package a
 julia> Pkg.add("Taro")
 ```
 
-On installation, the `tika-app-1.4.jar` file will be downloaded from *Maven Central*
-and `fop-2.0` will be downloaded from an Apache mirror.
+On installation, the `tika-app-1.17.jar` file will be downloaded from *Maven Central*
+and `fop-2.2` will be downloaded from an Apache mirror.
 
 ## Usage
 
@@ -31,4 +31,4 @@ This will set up the correct classpath, and initialise the JVM.
  using Taro
  Taro.init()
 ```
-Note: The reason why we do not run init() automatically on module load has to do with the fact that only one embedded JVM can be loaded per process. We need to set the classpath when we start the JVM. Thus, when we load two different packages which both depend on the JVM, we need provide the ability for all packages to modify the Java classpath. 
+Note: The reason why we do not run init() automatically on module load has to do with the fact that only one embedded JVM can be loaded per process. We need to set the classpath when we start the JVM. Thus, when we load two different packages which both depend on the JVM, we need provide the ability for all packages to modify the Java classpath.

src/Taro.jl

@@ -4,7 +4,7 @@ using JavaCall
 using DataFrames
 using DataArrays
 
-tika_jar = joinpath(dirname(@__FILE__), "..", "deps", "tika-app-1.10.jar")
+tika_jar = joinpath(dirname(@__FILE__), "..", "deps", "tika-app-1.17.jar")
 fop_lib = joinpath(dirname(@__FILE__), "..", "deps", "fop-2.0", "lib", "*")
 fop_jar = joinpath(dirname(@__FILE__), "..", "deps", "fop-2.0", "build", "fop.jar")