* When a BaseData instance has Nullable fields, the number of data
points per Series is inconsistent, and results in a Series with
a length different from the other Series we produce, resulting
in an error "ValueError: cannot handle a non-unique multi-index!"
when we were constructing the final DataFrame.
Use the original `pandas.Series` to create `Series` objects before `DataFrame` (wrapper version) creation.
It improves the speed because it avoids unnecessary and expensive index wrapping operations of the `DataFrame` creation.
In this implementation, we dynamically create new classes that wraps key functions and properties. The wrappers will map/convert any parameter that are convertible to the string representation of Symbol.ID before they are used by the original function/property.
Methods: 'diff', 'div', 'divide', 'drop', 'drop_duplicates', 'droplevel', 'dropna', 'dtypes', 'duplicated'
`BackwardsCompatibilityDataFrame_binary_operator` replaces `BackwardsCompatibilityDataFrame_add` to handle all operations (more to be added in future commits)
`Remapper.__getitem__` was not returns a `Remapper` object when the result was `pandas.DataFrame`. It is needed for a sequence of `.loc.` calls.
Refactors `Remapper._self_mapper` to handle tuples where the key can be found in both first and second position.
Update unit tests that should test `Symbol` object as key, but were using `str(Symbol)`.
- Add support for Symbol key access for pandas ix and iloc results
- Wrapp pdf merge, join, concat method results
- Wrapp pandas.concat method result with `Remapper`
- Adding unit tests
- Covers another level on inheritance of Market Data by using `Type.IsAssignableFrom`
- Caches the list of `MethodInfo` for custom data types to avoid redefining that list.
When custom data classes inherited from market data classes such as `TradeBar`, it created duplicate entries. Therefore, we need to exclude the common properties in the private field `PandasData._members`.
- Moving mapper from C# to Python since some cases did not work when
implemented in C#
- Small changes to `PandasDataFrameHistoryAlgorithm` which runs till the
end with no errors
- Adding more backwards compatible unit tests
- Add Pandas backwards compatibility shim
- Adding `MappingExtensions` which will remove data type from the
`Symbol.ID.Symbol` value to resolve the `MapFile`
- `SecurityIdentifier.TryParse()` will throw when given an invalid
`SecurityType`
Since Lean/QuantConnect data from a symbol can be found in Slice.Ticks, Slice.Bars and Slice.QuoteBars, information for all of this members must be used in order present all information in the pandas.DataFrame.
- We were excluding Exchange and Suspicious of quote ticks when they can be set with security types other than equity
- Custom data was not expecting non-numerical properties
- Moves PandasData class to its own file
- Refactors PandasData class to deal with list of tick
- Other minor changes requested by peer-review
- Adds unit test for PandasConverter's tick handling
Alexandre Catarino