Metadata-Version: 2.1
Name: bocks-ds
Version: 0.2.1
Summary: DataSource Python SDK
Home-page: https://github.com/bocks-ds/datasource-python
Author: Bocks
License: UNKNOWN
Platform: UNKNOWN
Classifier: Development Status :: 3 - Alpha
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: License :: Free for non-commercial use
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Requires-Dist: pycurl

# datasource-python
Python SDK for all DataSource iterations


## Installation

Install this package via pip:

```python
pip install bocks_ds
```


## Usage

### The Client

Import this package and instantiate a client object.

```python
import bocks_ds

ds_client = bocks_ds.Client("starfinder") # provide the name of the datasource to access
```

#### Error Example:

```python
try:
    bad_client = Client("bad_client")
except DSTargetError as e:
    print(e) # The target 'bad_client' provided in Client initialization is not in available target names:\n['starfinder', 'pathfinder']
```

### Fetching Data

There are currently only two options to consider when fetching data: the type and the parameters.

The "type" can be thought of as a data table, though we will discuss some complexities later on.


```python
response = ds_client.armor.get(['name', 'price']) # 'armor' here is the query type
if response.status_code == 200: # optional status check
    data = response.json() # '.json()' gives the API data output we came for
```

Note that the `ds_client` will accept **any value** as an attribute, which it will then use to craft an API request. 

#### Error Example:

```python
try:
    client.bad_name.get(['name']).json()
except DSQueryError as e:
    print(e) # <Response 400> DataSource did not find table/field 'bad_name'.
```

### Refining Your Request

The API allows for active filtering of strings and integers within your request. To do this, you'll need to set arguments immediately prior to making the request. 

For data values that return strings, you may refine with the terms `<value_name>_like` and `<value_name>_is`. For "like" queries, sequencing is not considered and the search is not case sensitive. For "is" queries, only exact, case-sensitive matches will be returned.

For data values that return integers, you may refine with the terms `<value_name>_min`, `<value_name>_max` and `<value_name>_equals`. These equate to "greater than or eqal to", "less than or equal to", and "equal to", respectively.

Finally, it is often valuable to select an item specifically by its `ID`, which functions in a more direct manner than the queries for integers (see example below).

All arguments must be presented as a single dictionary, as seen in the examples below.

```python
client.armor.set_arguments({"name_like":"basic"})
response = ds_client.armor.get(['name', 'price'])
```

```python
client.armor.set_arguments({"price_min":200, "price_max":2000})
response = ds_client.armor.get(['name', 'price'])
```

```python
client.armor.set_arguments({"id":1})
response = ds_client.armor.get(['name', 'price'])
```

Note that these requests do not stack, but you can place all terms into a single dictionary to futher refine results. If `set_arguments` is called multiple times, the final call will overwrite previous calls.

```python
query = {"name_like":"basic", "price_min":200, "price_max":2000}
client.armor.set_arguments(query)
response = ds_client.armor.get(['name', 'price'])
```

As a final note: all arguments are cleared when `get` is called.

#### Error Example

```python
try:
    client.armor.set_arguments({"name_min":200})
    erroneous_armor = client.armor.get(['name', 'price']) # Throws exception due to errors in response.json()
except DSQueryError as e:
    print(e) # <Response 400> ['Unknown argument "name_min" on field "Query.armor". Did you mean "name_is", "name_like", "type_min", "bulk_min", or "level_min"?']
```

### Fetching Nested Data

As mentioned previously, not all types match their data tables precisely. Some types include additional data from relationships with other tables. Documentation for these relationships is automatically generated in the API Documentation (such as at *docs.sfdatasource.com*).

In order to access nested data, it is required that you present a dictionary describing the data to fetch.

Where we previously provided a list of string field-names to query, we will now include a dictionary in that list, as seen in the examples below.

```python
query = [
    "name",
    {
        "effect_ranges": ["name", "description"],
    }
]
all_spells = ds_client.spells.get(query)
```

This logic is recursive, so in the event that a relationship target has it's own relationships you may do the following:

```python
query = [
    "name",
    {
        "class_proficiencies": ['name'],
        "class_features": ["name", {
            "class_special_skills": ['name']
        }],
    }
]
spells = client.classes.get(query)
```

