Add managed crontab support

README.md

@@ -41,17 +41,20 @@ intake
   },
   "env": {
     "...": "..."
-  }
+  },
+  "cron": "* * * * *"
 }
 ```
 
-Each key under `action` defines an action that can be taken for the source. `action` must be present with a `fetch` action. `env` is optional.
+Each key under `action` defines an action that can be taken for the source. A minimal `intake.json` must contain a `fetch` action with an `exe`. All other configs are optional.
+
+The `fetch` action is executed by `intake update`. All other actions are executable with `intake action`. `intake crontab` will create a crontab entry for each source with a `cron` config defined. The value of `cron` is the crontab spec according to which the source should be updated.
 
 ## Interface for source programs
 
-Intake interacts with sources by executing the actions defined in the source's `intake.json`. The `fetch` action is required and used to check for new feed items.
+Intake interacts with sources by executing the actions defined in the source's `intake.json`. The `fetch` action is required and used to check for new feed items when `intake update` is executed.
 
-When any action is executed, intake executes the `exe` program for the action with the corresponding `args` as arguments. The process's working directory is set to the source's folder, i.e. the folder containing `intake.json`. The process's environment is as follows:
+To execute an action, intake executes the `exe` program for the action with the corresponding `args` (if present) as arguments. The process's working directory is set to the source's folder, i.e. the folder containing `intake.json`. The process's environment is as follows:
 
 * intake's environment is inherited.
 * `STATE_PATH` is set to the absolute path of `state`.

demo/alice/currenttime/intake.json

@@ -4,5 +4,6 @@
       "exe": "currenttime.sh",
       "args": []
     }
-  }
+  },
+  "cron": "* * * * *"
 }

demo/default.nix

@@ -8,17 +8,16 @@
     isNormalUser = true;
     password = "alpha";
     uid = 1000;
+    packages = [ pkgs.intake ];
   };
 
   users.users.bob = {
     isNormalUser = true;
     password = "beta";
     uid  = 1001;
+    packages = [ pkgs.intake ];
   };
 
-  # Put intake on both users' PATH
-  environment.systemPackages = [ pkgs.intake ];
-
   # Set up intake for both users with an entry point at port 8080
   services.intake = {
     listen.port = 8080;

intake/app.py

@@ -7,7 +7,16 @@ import json
 import os
 import time
 
-from flask import Flask, render_template, request, jsonify, abort, redirect, url_for, current_app
+from flask import (
+    Flask,
+    render_template,
+    request,
+    jsonify,
+    abort,
+    redirect,
+    url_for,
+    current_app,
+)
 
 from intake.core import intake_data_dir
 from intake.source import LocalSource, execute_action, Item

intake/cli.py

@@ -11,11 +11,11 @@ import subprocess
 import sys
 
 from intake.core import intake_data_dir
+from intake.crontab import update_crontab_entries
 from intake.source import fetch_items, LocalSource, update_items, execute_action
 from intake.types import InvalidConfigException, SourceUpdateException
 
 
-
 def cmd_edit(cmd_args):
     """Open a source's config for editing."""
     parser = argparse.ArgumentParser(
@@ -296,6 +296,24 @@ def cmd_passwd(cmd_args):
     return 0
 
 
+def cmd_crontab(cmd_args):
+    """Update cron with configured source cron entries."""
+    parser = argparse.ArgumentParser(
+        prog="intake crontab",
+        description=cmd_crontab.__doc__,
+    )
+    parser.add_argument(
+        "--data",
+        "-d",
+        help="Path to the intake data directory containing source directories",
+    )
+    args = parser.parse_args(cmd_args)
+
+    data_path: Path = Path(args.data) if args.data else intake_data_dir()
+    update_crontab_entries(data_path)
+    return 0
+
+
 def cmd_run(cmd_args):
     """Run the default Flask server."""
     parser = argparse.ArgumentParser(

intake/crontab.py

@@ -0,0 +1,82 @@
+from pathlib import Path
+import os
+import subprocess
+
+from intake.source import LocalSource
+
+
+INTAKE_CRON_BEGIN = "### begin intake-managed crontab entries"
+INTAKE_CRON_END = "### end intake-managed crontab entries"
+
+
+def get_desired_crons(data_path: Path):
+    """
+    Get a list of sources and crontab specs from the data directory.
+    """
+    for child in data_path.iterdir():
+        if not (child / "intake.json").exists():
+            continue
+        source = LocalSource(data_path, child.name)
+        config = source.get_config()
+        if cron := config.get("cron"):
+            yield f"{cron}  . /etc/profile; intake update -s {source.source_name}"
+
+
+def update_crontab_entries(data_path: Path):
+    """
+    Update the intake-managed section of the user's crontab.
+    """
+    # If there is no crontab command available, quit early.
+    crontab_exists = subprocess.run(["command", "-v" "crontab"], shell=True)
+    if crontab_exists.returncode:
+        print("Could not update crontab")
+        return
+
+    # Get the current crontab
+    get_crontab = subprocess.run(
+        ["crontab", "-e"],
+        env={**os.environ, "EDITOR": "cat"},
+        stdout=subprocess.PIPE,
+        stderr=subprocess.DEVNULL,
+    )
+    crontab_lines = get_crontab.stdout.decode("utf-8").splitlines()
+
+    # Splice the intake crons into the crontab
+    new_crontab_lines = []
+    section_found = False
+    in_section = False
+    for i in range(len(crontab_lines)):
+
+        if not section_found and crontab_lines[i] == INTAKE_CRON_BEGIN:
+            section_found = True
+            in_section = True
+            # Open the section and add everything
+            new_crontab_lines.append(INTAKE_CRON_BEGIN)
+            new_crontab_lines.extend(get_desired_crons(data_path))
+
+        elif crontab_lines[i] == INTAKE_CRON_END:
+            new_crontab_lines.append(INTAKE_CRON_END)
+            in_section = False
+
+        elif not in_section:
+            new_crontab_lines.append(crontab_lines[i])
+
+    # If the splice mark was never found, append the whole section to the end
+    if not section_found:
+        new_crontab_lines.append(INTAKE_CRON_BEGIN)
+        new_crontab_lines.extend(get_desired_crons(data_path))
+        new_crontab_lines.append(INTAKE_CRON_END)
+
+    # Save the updated crontab
+    new_crontab: bytes = "\n".join(new_crontab_lines).encode("utf8")
+    save_crontab = subprocess.Popen(
+        ["crontab", "-"],
+        stdin=subprocess.PIPE,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.PIPE,
+    )
+    (stdout, stderr) = save_crontab.communicate(new_crontab)
+    for line in stdout.decode("utf8").splitlines():
+        print("[stdout]", line)
+    for line in stderr.decode("utf8").splitlines():
+        print("[stderr]", line)

module.nix

@@ -74,7 +74,7 @@ in {
     intakeDir = "/etc/intake";
     intakePwd = "${intakeDir}/htpasswd";
   in {
-    # Apply the overlay so intake is included inpkgs.
+    # Apply the overlay so intake is included in pkgs.
     nixpkgs.overlays = [ flake.overlays.default ];
 
     # Define a user group for access to the htpasswd file. nginx needs to be able to read it.
@@ -105,6 +105,9 @@ in {
       };
     in mkMerge (map addPackagesToUser enabledUserNames);
 
+    # Enable cron
+    services.cron.enable = true;
+
     # Define a user service for each configured user
     systemd.services =
     let