Added a stub for logging and long term storage.

main.py

@@ -2,12 +2,17 @@ import json
 import uuid
 import os
 import pathlib
-
+from uuid import uuid4
 from datetime import datetime
+import dataclasses
+
 from fasthtml.common import *
+from fastcore.all import typedispatch
 from sqlite_minutils.db import Database
+from huggingface_hub import CommitScheduler
 
 import eval_code
+import storage
 
 # the secrets. Will be loaded from HF, or for docker --env-file or from IDE
 
@@ -15,16 +20,40 @@ OAUTH_CLIENT_ID = os.environ.get('OAUTH_CLIENT_ID')
 OAUTH_SCOPES = os.environ.get('OAUTH_SCOPES')
 OAUTH_CLIENT_SECRET = os.environ.get('OAUTH_CLIENT_SECRET')
 OPENID_PROVIDER_URL = os.environ.get('OPENID_PROVIDER_URL')
-SPACE_HOST = os.environ.get('SPACE_HOST')
+SPACE_HOST = os.environ.get('SPACE_HOST', "none")
+HF_DATASET_AUTH_TOKEN = os.environ.get('HF_DATASET_AUTH_TOKEN')
+
+assert SPACE_HOST is not "none", "Please set the corrent ENV variables!!!"
 
-# The database. In memory bc it is efemerial on HF anyway.
+LOCAL_STORAGE_PATH = ""
 
+FILE_EVENTS = f"events-{datetime.utcnow()}-{uuid4()}.json"
 
 if "localhost" in SPACE_HOST:
     DATABASE_NAME = "data/sessions_meta.db"
+    LOCAL_STORAGE_PATH = Path("data/persistent/")
     pathlib.Path(DATABASE_NAME).unlink(missing_ok=True)
 else:
-    DATABASE_NAME = "/data/sessions_meta.db"
+    DATABASE_NAME = "/tmp/cache/sessions_meta.db"
+    LOCAL_STORAGE_PATH = Path("/tmp/cache/persistent")
+
+
+LOCAL_STORAGE_PATH.mkdir(exist_ok=True, parents=True)
+EVENTS_FILE_PATH = LOCAL_STORAGE_PATH / FILE_EVENTS
+
+
+scheduler = CommitScheduler(
+    repo_id="ml-visoft/c-reviewer",
+    repo_type="dataset",
+    folder_path=LOCAL_STORAGE_PATH,
+    every=100,
+    path_in_repo="raw_data",
+    token=HF_DATASET_AUTH_TOKEN,
+    allow_patterns=".jslines",
+    squash_history=False
+)
+
+
 
 global_database = Database(DATABASE_NAME)
 global_database_tables = global_database.t
@@ -74,6 +103,19 @@ else:
     REFRESH_TIME = 1
 
 
+def untyped_save_to_storage(dc, filename):
+    with scheduler.lock:
+        js_str = json.dumps(dataclasses.asdict(dc)) + "\n"
+        with open(filename, "a") as f:
+            f.write(js_str)
+
+
+@typedispatch
+def save_to_storage(nav_event:storage.NavigationEvent):
+    untyped_save_to_storage(nav_event, EVENTS_FILE_PATH)
+
+
+
 def vallidate_and_get_question_evaluation_objectid(session, qe_id:int):
     if 'session_id' not in session: return None
     session_id = session["session_id"]
@@ -93,7 +135,7 @@ def vallidate_and_get_question_evaluation_objectid(session, qe_id:int):
     return True, qe_obj
 
 
-def html_create_feedback_updown_button(qe_id, ans_id, selected=0, disabled=0):
+def html_create_feedback_updown_button(qe_id, ans_id, selected=0, disabled=False):
     html_target_id = f"buttons_{ans_id}"
     colors = ["grey", "blue"]
     up_col = colors[0]
@@ -151,6 +193,11 @@ def put(session, qe_id:int, ans_id:int, which:int):
     if qe_id != qe_obj.id:
         print(f"QE {qe_id} does not belong to {qe_obj.id}")
         return None
+
+    # save_to_storage(
+    #     storage.NavigationEvent(event_type="/", event_session_id=session_id, event_params={"qe_id":qe_id})
+    # )
+
     answer_eval_js = json.loads(qe_obj.answer_eval_text)
     crt_selection = answer_eval_js[ans_id]["EVAL"]
     input_button = which
@@ -179,8 +226,12 @@ def html_get_textual_feedback_form(qe_obj, thank=False):
 
 @rt("/submit_feedback/{qe_id}")
 def post(session, qe_id:int, freeform_feedback:str):
-    print(qe_id, freeform_feedback)
     # Update the object
+    session_id = session.get("session_id", "Not set")
+    save_to_storage(
+        storage.NavigationEvent(event_type="/submit_feedback", event_session_id=session_id,
+                                event_params={"qe_id":qe_id})
+    )
     is_ok, qe_obj = vallidate_and_get_question_evaluation_objectid(session, qe_id)
     if not is_ok:
         return "Error"
@@ -350,6 +401,9 @@ def get(session):
     if 'session_id' not in session:
         session['session_id'] = str(uuid.uuid4())
     session_id = session["session_id"]
+    save_to_storage(
+        storage.NavigationEvent(event_type="/", event_session_id=session_id)
+    )
     title = Title('C code review for students')
     preamble = [H1("Evaluate your C code!"),
                 P("Enter your code in the textbox below and wait for answers."),
@@ -424,6 +478,10 @@ def post(ccodetoeval:str, session):
 def get(session):
     if 'session_id' not in session: return P("Bad call. No session ID")
     session_id = session["session_id"]
+    save_to_storage(
+        storage.NavigationEvent(event_type="/clear_area", event_session_id=session_id)
+    )
+
     # insert a row to "cancel"/reset the current request
     session_obj = Session_State_cls(
         session_id=session_id,

requirements.txt

@@ -1,2 +1,3 @@
 python-fasthtml
-fastsql==1.0.1
+fastsql==1.0.1
+huggingface_hub==-0.24.5

storage.py

@@ -0,0 +1,12 @@
+import dataclasses
+from datetime import datetime
+
+
+@dataclasses.dataclass(slots=True)
+class NavigationEvent():
+    version:int = 1
+    event_type:str = '(none)'
+    event_date:str = datetime.isoformat(datetime.utcnow())
+    event_session_id:str = ""
+    event_params:dict=dataclasses.field(default_factory=dict)
+