Merge branch 'main' into sort-results-popularity

README.md

@@ -4,7 +4,7 @@
 
 ### Prerequisites:
 
-1. Python (≥ `3.10`) installed on your system.
+1. Python 3.12.3 installed on your system.
 2. Ensure you have `poetry` installed. If not, you can install them using `pip`.
 
 ```bash
@@ -135,4 +135,15 @@ The [French regional TMDB Series Dataset](https://huggingface.co/datasets/DataFo
 
 ```bash
 invoke update-series-dataset
-```
+```
+
+## Python CLI
+
+The Python CLI supports the following commands:
+
+```
+python -m observatoire.tmdb.movies --mode=[latest | missing]
+python -m observatoire.tmdb.series --mode=[latest | missing]
+```
+
+In the `latest` mode, which is the default, these commands sync the latest records from TMDB to our datasets on Hugging Face.  In the `missing` mode, they calculate which rows may be missing from the Hugging Face datasets and attempt to sync these records.

observatoire/tmdb/movies/__main__.py

@@ -2,6 +2,7 @@
 import contextlib
 import os
 
+import fire
 from tqdm import tqdm
 
 from observatoire.tmdb.config import HF_MOVIES_DATASET, TMDB_BATCH_SIZE
@@ -10,32 +11,55 @@
 from observatoire.tmdb.logger import setup_logger
 from observatoire.tmdb.movies.data import make_movie_df
 from observatoire.tmdb.movies.tmdb import get_latest_movie_id, get_movie_data
+from observatoire.tmdb.types import Mode
 
 
-def executor() -> None:
+def executor(mode: Mode = "latest") -> None:
     logger = setup_logger()
-    logger.info("Starting Executor")
-
-    # first, lets get the id of the lastest movies
-    latest_id = get_latest_movie_id()
+    logger.info("Starting Movie Dataset Update")
 
     # second, let's get the last movie from our last run
     df_current = load_dataset(HF_MOVIES_DATASET)
     current_id = df_current["id"].max() if df_current is not None else None
 
-    # Generate a list of movie IDs
-    movie_ids_list = list(range(current_id or 1, latest_id))
-    total_movies_to_process = len(movie_ids_list)
+    if mode == "missing":
+        logger.info("Mode: Update Missing Movies")
+
+        # Create a set of all possible ids
+        all_ids = set(range(1, current_id + 1))
+
+        # Create a set of ids present in df_current
+        current_ids = set(df_current["id"].dropna())
+
+        # Find the difference between all_ids and current_ids
+        missing_ids = all_ids - current_ids
+
+        # Convert the set to a list
+        movie_ids_list = list(missing_ids)
+
+        logger.info(f"Total Missing Movies: {len(movie_ids_list)}")
+    elif mode == "latest":
+        logger.info("Mode: Update Latest MOvies")
+
+        # first, lets get the id of the lastest movie
+        latest_id = get_latest_movie_id()
+
+        # Generate a list of movie IDs
+        movie_ids_list = list(range(current_id or 1, latest_id))
+
+        logger.info(f"Total Latest Movies: {len(movie_ids_list)}")
 
-    logger.info(f"Total Movies to Process in this run: {total_movies_to_process}")
+    else:
+        logger.critical("Invalid mode selected")
+        return
 
     # Split movie_ids_list into chunks of TMDB_BATCH_SIZE
     batches = [
         movie_ids_list[i : i + TMDB_BATCH_SIZE]
         for i in range(0, len(movie_ids_list), TMDB_BATCH_SIZE)
     ]
 
-    with tqdm(total=total_movies_to_process, unit=" movies") as pbar:
+    with tqdm(total=len(movie_ids_list), unit=" movies") as pbar:
         for batch in batches:
             logger.info(f"Processing batch of {len(batch)} movies")
 
@@ -75,4 +99,4 @@ def executor() -> None:
 
 
 if __name__ == "__main__":
-    executor()
+    fire.Fire(executor)

observatoire/tmdb/series/__main__.py

@@ -2,6 +2,7 @@
 import contextlib
 import os
 
+import fire
 from tqdm import tqdm
 
 from observatoire.tmdb.config import HF_SERIES_DATASET, TMDB_BATCH_SIZE
@@ -10,32 +11,55 @@
 from observatoire.tmdb.logger import setup_logger
 from observatoire.tmdb.series.data import make_series_df
 from observatoire.tmdb.series.tmdb import get_latest_series_id, get_series_data
+from observatoire.tmdb.types import Mode
 
 
-def executor() -> None:
+def executor(mode: Mode = "latest") -> None:
     logger = setup_logger()
-    logger.info("Starting Executor")
-
-    # first, lets get the id of the lastest series
-    latest_id = get_latest_series_id()
+    logger.info("Starting Series Dataset Update")
 
     # second, let's get the last series from our last run
     df_current = load_dataset(HF_SERIES_DATASET)
     current_id = df_current["id"].max() if df_current is not None else None
 
-    # Generate a list of series IDs
-    series_ids_list = list(range(current_id or 1, latest_id))
-    total_series_to_process = len(series_ids_list)
+    if mode == "missing":
+        logger.info("Mode: Update Missing Series")
+
+        # Create a set of all possible ids
+        all_ids = set(range(1, current_id + 1))
+
+        # Create a set of ids present in df_current
+        current_ids = set(df_current["id"].dropna())
+
+        # Find the difference between all_ids and current_ids
+        missing_ids = all_ids - current_ids
+
+        # Convert the set to a list
+        series_ids_list = list(missing_ids)
+
+        logger.info(f"Total Missing Series: {len(series_ids_list)}")
+    elif mode == "latest":
+        logger.info("Mode: Update Latest Series")
+
+        # first, lets get the id of the lastest series
+        latest_id = get_latest_series_id()
+
+        # Generate a list of series IDs
+        series_ids_list = list(range(current_id or 1, latest_id))
+
+        logger.info(f"Total Latest Series: {len(series_ids_list)}")
 
-    logger.info(f"Total Series to Process in this run: {total_series_to_process}")
+    else:
+        logger.critical("Invalid mode selected")
+        return
 
     # Split series_ids_list into chunks of TMDB_BATCH_SIZE
     batches = [
         series_ids_list[i : i + TMDB_BATCH_SIZE]
         for i in range(0, len(series_ids_list), TMDB_BATCH_SIZE)
     ]
 
-    with tqdm(total=total_series_to_process, unit=" series") as pbar:
+    with tqdm(total=len(series_ids_list), unit=" series") as pbar:
         for batch in batches:
             logger.info(f"Processing batch of {len(batch)} series")
 
@@ -75,4 +99,4 @@ def executor() -> None:
 
 
 if __name__ == "__main__":
-    executor()
+    fire.Fire(executor)

observatoire/tmdb/series/data.py

@@ -1,5 +1,4 @@
 import json
-import sys
 
 import pandas as pd
 
@@ -12,12 +11,14 @@
     safe_list,
     safe_str,
 )
+from observatoire.tmdb.logger import setup_logger
 
 
 def make_series_df(series_json: list[str]) -> pd.DataFrame:
     """
     Transforms the JSON data into a DataFrame
     """
+    logger = setup_logger()
 
     data = []
     unique_ids = set()
@@ -82,8 +83,8 @@ def make_series_df(series_json: list[str]) -> pd.DataFrame:
             safe_data["vote_average"] = safe_float(line_in_json, "vote_average")
             safe_data["vote_count"] = safe_int(line_in_json, "vote_count")
 
-        except Exception as e:
-            print(f"Error: {e}", file=sys.stderr)
+        except Exception:
+            logger.exception(f"Could not parse line: {line_in_json}")
             continue
 
         # remove and newline chracters

observatoire/tmdb/types.py

@@ -0,0 +1,3 @@
+from typing import Literal
+
+Mode = Literal["latest", "missing"]

pyproject.toml

@@ -26,6 +26,7 @@ bs4 = "^0.0.2"
 python-dotenv = "^1.0.1"
 datasets = "^2.18.0"
 pycountry-convert = "^0.7.2"
+fire = "^0.6.0"
 
 [tool.poetry.group.dev.dependencies]
 pre-commit = "^2.20.0"