cache & refactor updates (ExpDev07#310)

* reduce cache size (#21) * formatting (120 length -> 100) * csbs to/from Redis * nyt to/from Redis * ignore locustfile * unused coordinates * cache redis json serialization error * fix nyt redis serialization error * refactor Timeline class to pydantic model * partial recovery fix (#31) * fix jhu timeline init call * update requirements * update isort usage Co-authored-by: DeepSource Bot <[email protected]> Co-authored-by: Thanh Nguyen <[email protected]> Co-authored-by: codedawi <[email protected]> Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com> Co-authored-by: Sourcery AI <[email protected]>
Kilo59 · Oct 12, 2020 · 69c2d02 · 69c2d02
1 parent 1508dce
commit 69c2d02
Show file tree

Hide file tree

Showing 25 changed files with 504 additions and 388 deletions.
diff --git a/.deepsource.toml b/.deepsource.toml
@@ -0,0 +1,10 @@
+version = 1
+
+test_patterns = ["tests/**"]
+
+[[analyzers]]
+name = "python"
+enabled = true
+
+  [analyzers.meta]
+  runtime_version = "3.x.x"
diff --git a/.gitignore b/.gitignore
@@ -51,6 +51,7 @@ htmlcov/
 nosetests.xml
 coverage.xml
 *,cover
+locustfile.py
 
 # Translations
 *.mo

diff --git a/Pipfile.lock b/Pipfile.lock
diff --git a/app/data/__init__.py b/app/data/__init__.py
@@ -4,7 +4,11 @@
 from ..services.location.nyt import NYTLocationService
 
 # Mapping of services to data-sources.
-DATA_SOURCES = {"jhu": JhuLocationService(), "csbs": CSBSLocationService(), "nyt": NYTLocationService()}
+DATA_SOURCES = {
+    "jhu": JhuLocationService(),
+    "csbs": CSBSLocationService(),
+    "nyt": NYTLocationService(),
+}
 
 
 def data_source(source):

diff --git a/app/io.py b/app/io.py
@@ -10,7 +10,11 @@
 
 
 def save(
-    name: str, content: Union[str, Dict, List], write_mode: str = "w", indent: int = 2, **json_dumps_kwargs
+    name: str,
+    content: Union[str, Dict, List],
+    write_mode: str = "w",
+    indent: int = 2,
+    **json_dumps_kwargs,
 ) -> pathlib.Path:
     """Save content to a file. If content is a dictionary, use json.dumps()."""
     path = DATA / name
@@ -35,7 +39,12 @@ class AIO:
 
     @classmethod
     async def save(
-        cls, name: str, content: Union[str, Dict, List], write_mode: str = "w", indent: int = 2, **json_dumps_kwargs
+        cls,
+        name: str,
+        content: Union[str, Dict, List],
+        write_mode: str = "w",
+        indent: int = 2,
+        **json_dumps_kwargs,
     ):
         """Save content to a file. If content is a dictionary, use json.dumps()."""
         path = DATA / name

diff --git a/app/location/__init__.py b/app/location/__init__.py
@@ -11,7 +11,7 @@ class Location:  # pylint: disable=too-many-instance-attributes
     """
 
     def __init__(
-        self, id, country, province, coordinates, last_updated, confirmed, deaths, recovered
+        self, id, country, province, coordinates, last_updated, confirmed, deaths, recovered,
     ):  # pylint: disable=too-many-arguments
         # General info.
         self.id = id
@@ -66,7 +66,11 @@ def serialize(self):
             # Last updated.
             "last_updated": self.last_updated,
             # Latest data (statistics).
-            "latest": {"confirmed": self.confirmed, "deaths": self.deaths, "recovered": self.recovered},
+            "latest": {
+                "confirmed": self.confirmed,
+                "deaths": self.deaths,
+                "recovered": self.recovered,
+            },
         }
 
 

diff --git a/app/main.py b/app/main.py
@@ -6,11 +6,10 @@
 import pydantic
 import sentry_sdk
 import uvicorn
-from fastapi import FastAPI, Request, Response, openapi
+from fastapi import FastAPI, Request, Response
 from fastapi.middleware.cors import CORSMiddleware
 from fastapi.middleware.gzip import GZipMiddleware
 from fastapi.responses import JSONResponse
-from fastapi.staticfiles import StaticFiles
 from scout_apm.async_.starlette import ScoutMiddleware
 from sentry_sdk.integrations.asgi import SentryAsgiMiddleware
 
@@ -35,9 +34,9 @@
         "API for tracking the global coronavirus (COVID-19, SARS-CoV-2) outbreak."
         " Project page: https://github.com/ExpDev07/coronavirus-tracker-api."
     ),
-    version="2.0.3",
-    docs_url=None,
-    redoc_url=None,
+    version="2.0.4",
+    docs_url="/",
+    redoc_url="/docs",
     on_startup=[setup_client_session],
     on_shutdown=[teardown_client_session],
 )
@@ -60,7 +59,11 @@
 
 # Enable CORS.
 APP.add_middleware(
-    CORSMiddleware, allow_credentials=True, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"],
+    CORSMiddleware,
+    allow_credentials=True,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
 )
 APP.add_middleware(GZipMiddleware, minimum_size=1000)
 
@@ -109,31 +112,6 @@ async def handle_validation_error(
 # Include routers.
 APP.include_router(V1, prefix="", tags=["v1"])
 APP.include_router(V2, prefix="/v2", tags=["v2"])
-APP.mount("/static", StaticFiles(directory="static"), name="static")
-
-# ##############
-# Swagger/Redocs
-# ##############
-
-
-@APP.get("/", include_in_schema=False)
-async def custom_swagger_ui_html():
-    """Serve Swagger UI."""
-    return openapi.docs.get_swagger_ui_html(
-        openapi_url=APP.openapi_url,
-        title=f"{APP.title} - Swagger UI",
-        oauth2_redirect_url=APP.swagger_ui_oauth2_redirect_url,
-        swagger_js_url="/static/swagger-ui-bundle.js",
-        swagger_css_url="/static/swagger-ui.css",
-    )
-
-
-@APP.get("/docs", include_in_schema=False)
-async def redoc_html():
-    """Serve ReDoc UI."""
-    return openapi.docs.get_redoc_html(
-        openapi_url=APP.openapi_url, title=f"{APP.title} - ReDoc", redoc_js_url="/static/redoc.standalone.js",
-    )
 
 
 # Running of app.

diff --git a/app/models.py b/app/models.py
@@ -1,7 +1,7 @@
 """app.models.py"""
 from typing import Dict, List
 
-from pydantic import BaseModel
+from pydantic import BaseModel, validator
 
 
 class Latest(BaseModel):
@@ -27,9 +27,26 @@ class Timeline(BaseModel):
     Timeline model.
     """
 
-    latest: int
     timeline: Dict[str, int] = {}
 
+    @validator("timeline")
+    @classmethod
+    def sort_timeline(cls, value):
+        """Sort the timeline history before inserting into the model"""
+        return dict(sorted(value.items()))
+
+    @property
+    def latest(self):
+        """Get latest available history value."""
+        return list(self.timeline.values())[-1] if self.timeline else 0
+
+    def serialize(self):
+        """
+        Serialize the model into dict
+        TODO: override dict() instead of using serialize
+        """
+        return {**self.dict(), "latest": self.latest}
+
 
 class Timelines(BaseModel):
     """

diff --git a/app/routers/v1.py b/app/routers/v1.py
@@ -19,7 +19,11 @@ async def all_categories():
         "deaths": deaths,
         "recovered": recovered,
         # Latest.
-        "latest": {"confirmed": confirmed["latest"], "deaths": deaths["latest"], "recovered": recovered["latest"],},
+        "latest": {
+            "confirmed": confirmed["latest"],
+            "deaths": deaths["latest"],
+            "recovered": recovered["latest"],
+        },
     }
 
 

diff --git a/app/routers/v2.py b/app/routers/v2.py
@@ -65,11 +65,17 @@ async def get_locations(
 
         # Do filtering.
         try:
-            locations = [location for location in locations if str(getattr(location, key)).lower() == str(value)]
+            locations = [
+                location
+                for location in locations
+                if str(getattr(location, key)).lower() == str(value)
+            ]
         except AttributeError:
             pass
         if not locations:
-            raise HTTPException(404, detail=f"Source `{source}` does not have the desired location data.")
+            raise HTTPException(
+                404, detail=f"Source `{source}` does not have the desired location data.",
+            )
 
     # Return final serialized data.
     return {
@@ -84,7 +90,9 @@ async def get_locations(
 
 # pylint: disable=invalid-name
 @V2.get("/locations/{id}", response_model=LocationResponse)
-async def get_location_by_id(request: Request, id: int, source: Sources = "jhu", timelines: bool = True):
+async def get_location_by_id(
+    request: Request, id: int, source: Sources = "jhu", timelines: bool = True
+):
     """
     Getting specific location by id.
     """

diff --git a/app/services/location/csbs.py b/app/services/location/csbs.py
@@ -6,6 +6,7 @@
 from asyncache import cached
 from cachetools import TTLCache
 
+from ...caches import check_cache, load_cache
 from ...coordinates import Coordinates
 from ...location.csbs import CSBSLocation
 from ...utils import httputils
@@ -34,7 +35,7 @@ async def get(self, loc_id):  # pylint: disable=arguments-differ
 BASE_URL = "https://facts.csbs.org/covid-19/covid19_county.csv"
 
 
-@cached(cache=TTLCache(maxsize=1, ttl=3600))
+@cached(cache=TTLCache(maxsize=1, ttl=1800))
 async def get_locations():
     """
     Retrieves county locations; locations are cached for 1 hour
@@ -44,48 +45,58 @@ async def get_locations():
     """
     data_id = "csbs.locations"
     LOGGER.info(f"{data_id} Requesting data...")
-    async with httputils.CLIENT_SESSION.get(BASE_URL) as response:
-        text = await response.text()
-
-    LOGGER.debug(f"{data_id} Data received")
-
-    data = list(csv.DictReader(text.splitlines()))
-    LOGGER.debug(f"{data_id} CSV parsed")
-
-    locations = []
-
-    for i, item in enumerate(data):
-        # General info.
-        state = item["State Name"]
-        county = item["County Name"]
-
-        # Ensure country is specified.
-        if county in {"Unassigned", "Unknown"}:
-            continue
-
-        # Coordinates.
-        coordinates = Coordinates(item["Latitude"], item["Longitude"])  # pylint: disable=unused-variable
-
-        # Date string without "EDT" at end.
-        last_update = " ".join(item["Last Update"].split(" ")[0:2])
-
-        # Append to locations.
-        locations.append(
-            CSBSLocation(
-                # General info.
-                i,
-                state,
-                county,
-                # Coordinates.
-                Coordinates(item["Latitude"], item["Longitude"]),
-                # Last update (parse as ISO).
-                datetime.strptime(last_update, "%Y-%m-%d %H:%M").isoformat() + "Z",
-                # Statistics.
-                int(item["Confirmed"] or 0),
-                int(item["Death"] or 0),
+    # check shared cache
+    cache_results = await check_cache(data_id)
+    if cache_results:
+        LOGGER.info(f"{data_id} using shared cache results")
+        locations = cache_results
+    else:
+        LOGGER.info(f"{data_id} shared cache empty")
+        async with httputils.CLIENT_SESSION.get(BASE_URL) as response:
+            text = await response.text()
+
+        LOGGER.debug(f"{data_id} Data received")
+
+        data = list(csv.DictReader(text.splitlines()))
+        LOGGER.debug(f"{data_id} CSV parsed")
+
+        locations = []
+
+        for i, item in enumerate(data):
+            # General info.
+            state = item["State Name"]
+            county = item["County Name"]
+
+            # Ensure country is specified.
+            if county in {"Unassigned", "Unknown"}:
+                continue
+
+            # Date string without "EDT" at end.
+            last_update = " ".join(item["Last Update"].split(" ")[0:2])
+
+            # Append to locations.
+            locations.append(
+                CSBSLocation(
+                    # General info.
+                    i,
+                    state,
+                    county,
+                    # Coordinates.
+                    Coordinates(item["Latitude"], item["Longitude"]),
+                    # Last update (parse as ISO).
+                    datetime.strptime(last_update, "%Y-%m-%d %H:%M").isoformat() + "Z",
+                    # Statistics.
+                    int(item["Confirmed"] or 0),
+                    int(item["Death"] or 0),
+                )
             )
-        )
-    LOGGER.info(f"{data_id} Data normalized")
+        LOGGER.info(f"{data_id} Data normalized")
+        # save the results to distributed cache
+        # TODO: fix json serialization
+        try:
+            await load_cache(data_id, locations)
+        except TypeError as type_err:
+            LOGGER.error(type_err)
 
     # Return the locations.
     return locations