[SPARK-11292][SQL] Python API for text data source

[rxin]

Adds DataFrameReader.text and DataFrameWriter.text.

[SparkQA]

Test build #44278 has finished for PR 9259 at commit f0e7201.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #44282 has finished for PR 9259 at commit ce990e9.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #44287 has finished for PR 9259 at commit ac7ea41.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):\n * class DefaultSource extends HadoopFsRelationProvider with DataSourceRegister\n * class TextOutputWriter(path: String, dataSchema: StructType, context: TaskAttemptContext)\n

[SparkQA]

Test build #44289 has finished for PR 9259 at commit b6ecf90.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[rxin]

cc @davies

[davies]

Should we also support multiple files or directories, as for parquet?

[rxin]

Could be a good idea. If we do that maybe we should do it for all data sources though, and maybe not part of this?

[yhuai]

merging to master.