draft of changes to support deployment on OpenLambda

.zenodo.json

@@ -0,0 +1,56 @@
+{
+  "title": "Ag Forecasting API",
+  "upload_type": "software",
+  "description": "The purpose of this API is to serve as the backend of crop disease forecasting inferences based on real-time weather data from Wisconet weather stations.",
+  "creators": [
+    {
+      "name": "Oros, Maria",
+      "affiliation": "UW-Madison",
+      "orcid": "0000-0009-0006-2067-4063",
+      "type": "ProjectMember"
+    },
+    {
+      "name": "Smith, Damon",
+      "affiliation": "UW-Madison",
+      "orcid": "0000-0003-3436-3718",
+      "type": "ProjectMember"
+    },
+    {
+      "name": "McConell, Iain",
+      "affiliation": "UW-Madison",
+      "orcid": "0000-0002-9982-7088",
+      "type": "ProjectMember"
+    }
+  ],
+  "contributors": [
+    {
+      "name": "University of Wisconsin - Data Science Institute",
+      "affiliation": "University of Wisconsin - Madison",
+      "type": "HostingInstitution"
+    }
+  ],
+  "access_right": "open",
+  "license": "mit",
+  "keywords": [
+    "data science",
+    "forecasting",
+    "open source software",
+    "agriculture",
+    "crop disease",
+    "Wisconet",
+    "UW-DSI"
+  ],
+  "version": "1.0.0",
+  "language": "eng",
+  "communities": [
+    { "identifier": "university-of-wisconsin-data-science-institute" }
+  ],
+  "related_identifiers": [
+    {
+      "identifier": "https://github.com/UW-Madison-DSI/ag_forecasting_api",
+      "relation": "isSupplementTo",
+      "resource_type": "software"
+    }
+  ],
+  "grants": []
+}

CITATION.cff

@@ -0,0 +1,28 @@
+cff-version: 1.2.0
+message: "Please cite this work when using the software."
+type: software
+title: "Soybean ROI tool"
+authors:
+  - family-names: "Oros"
+    given-names: "Maria"
+    orcid: "https://orcid.org/0000-0009-0006-2067-4063"
+    affiliation: "UW-Madison"
+  - family-names: "Smith"
+    given-names: "Damon"
+    orcid: "https://orcid.org/0000-0003-3436-3718"
+    affiliation: "UW-Madison"
+  - family-names: "McConnel"
+    given-names: "Iain"
+    orcid: "https://orcid.org/0000-0002-9982-7088"
+    affiliation: "UW-Madison"
+repository-code: "https://github.com/UW-Madison-DSI/ag_forecasting_api"
+url: "https://github.com/UW-Madison-DSI/ag_forecasting_api"
+license: "MIT"
+keywords:
+  - open source
+  - forecasting-tools
+  - fungicide
+  - agriculture
+# recommended:
+version: "0.1.0"
+date-released: "2025-12-15"

README.md

@@ -1,8 +1,13 @@
 # Open Source Agricultural Forecasting API
 
+## Citation
+
+[![DOI](https://zenodo.org/badge/837313087.svg)](https://doi.org/10.5281/zenodo.17459876)
+
+## About
+
 This API provides access to crop disease models developed by the University of wisconesin Madison experts in plant pathology. The API uses a FastAPI-based backend and integrates weather station data from public and private sources, including mesonet weather stations and IBM Environmental Intelligence. It also supports querying Wisconet by a wrapper built in top of it in a custom way. 
 [API](https://ag-forecasting-api.services.dsi.wisc.edu/)
-#(https://connect.doit.wisc.edu/pywisconet_wrapper/docs#/default/all_data_from_wisconet_query_ag_models_wrappers_wisconet_get)
 
 Table of Contents
 

ag_models_wrappers/process_wisconet.py

@@ -8,7 +8,7 @@
 import pickle
 import traceback
 from datetime import datetime, timedelta
-from concurrent.futures import ProcessPoolExecutor, as_completed
+from concurrent.futures import ThreadPoolExecutor, as_completed
 from ag_models_wrappers.forecasting_models import (
     calculate_tarspot_risk_function,
     calculate_gray_leaf_spot_risk_function,
@@ -25,10 +25,10 @@
 BASE_URL = "https://wisconet.wisc.edu/api/v1"
 MIN_DAYS_ACTIVE = 38
 
-MEASUREMENTS_CACHE_DIR = "station_measurements_cache"
+MEASUREMENTS_CACHE_DIR = os.getenv("MEASUREMENTS_CACHE_DIR", "station_measurements_cache")
 os.makedirs(MEASUREMENTS_CACHE_DIR, exist_ok=True)
 
-STATIONS_CACHE_FILE = "wisconsin_stations_cache.csv"
+STATIONS_CACHE_FILE = os.getenv("STATIONS_CACHE_FILE", "wisconsin_stations_cache.csv")
 CACHE_EXPIRY_DAYS = 7
 STATIONS_TO_EXCLUDE = ['MITEST1', 'WNTEST1']
 
@@ -286,7 +286,7 @@ async def retrieve_tarspot_all_stations_async(input_date, input_station_id=None,
             # 1) Load or fetch station list, with caching
             stations = None
 
-            if today.day == 1:
+            if today.day == 1 or not os.path.exists(STATIONS_CACHE_FILE):
                 url = f"https://api.wisconet.wisc.edu/api/v1/stations/"
                 async with session.get(url) as resp:
                     if resp.status == 200:
@@ -303,7 +303,7 @@ async def retrieve_tarspot_all_stations_async(input_date, input_station_id=None,
                         stations.to_csv(STATIONS_CACHE_FILE, index=False)
                     else:
                         stations= None
-            if today.day!=1 or stations==None:
+            if stations is None:
                 stations = pd.read_csv(STATIONS_CACHE_FILE)
 
             station_ids = stations['station_id'].tolist()
@@ -325,7 +325,7 @@ async def retrieve_tarspot_all_stations_async(input_date, input_station_id=None,
 
             # 5) Compute risks in parallel
             chunks = chunk_dataframe(merged, os.cpu_count() or 1)
-            with ProcessPoolExecutor() as exe:
+            with ThreadPoolExecutor() as exe:
                 futures = [exe.submit(compute_risks, c) for c in chunks]
                 processed = [f.result() for f in as_completed(futures)]
             final = pd.concat(processed, ignore_index=True)