casaconfig 1.3.1__py3-none-any.whl → 1.4.0__py3-none-any.whl

casaconfig/private/measures_update.py

@@ -15,11 +15,12 @@
 this module will be included in the api
 """
 
-def measures_update(path=None, version=None, force=False, logger=None, auto_update_rules=False, use_astron_obs_table=False, verbose=None):
+def measures_update(path=None, version=None, force=False, measures_site=None, logger=None, auto_update_rules=False, use_astron_obs_table=False, verbose=None):
     """
-    Update or install the IERS data used for measures calculations from ASTRON into path.
+    Update or install the IERS data used for measures calculations from measures_site into path.
     
-    Original data source used by ASTRON is here: https://www.iers.org/IERS/EN/DataProducts/data.html
+    Original IERS data source used by each of the recommended measures sites
+    is here: https://www.iers.org/IERS/EN/DataProducts/data.html
 
     If no update is necessary then this function will silently return.
 
@@ -27,16 +28,16 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
     are unchanged for expected reasons. A level of 0 outputs nothing. A
     value of 1 sends any output to the logger and a value of 2 logs and prints the information.
     The default value of the verbose argument is taken from the casaconfig_verbose config
-    value (defaults to 1).
+    value (defaults to 1). Error messages are always logged and printed.
 
     CASA maintains a separate Observatories table which is available in the casarundata
-    collection through pull_data and data_update. The Observatories table found at ASTRON
+    collection through pull_data and data_update. The Observatories table found at measures_site
     is not installed by measures_update and any Observatories file at path will not be changed
     by using this function. This behavior can be changed by setting force and use_astron_obs_table
     both to True (use_astron_obs_table is ignored when force is False).
 
-    A text file (readme.txt in the geodetic directory in path) records the measures version string
-    and the date when that version was installed in path.
+    A text file (readme.txt in the geodetic directory in path) records the measures version string,
+    the measures site that was used and the date when that version was installed in path.
 
     If path is None then config.measurespath is used.
 
@@ -44,12 +45,26 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
     nothing unless force is True.
 
     If a specific version is not requested (the default) and the modification time of that text
-    file is less than 24 hrs before now then this function does nothing unless force is True. When this
-    function checks for a more recent version and finds that the installed version is the most recent
-    then the modification time of that text file is changed to the current time even though nothing has
-    changed in path. This limits the number of attempts to update the measures data (including checking
-    for more recent data) to once per day. When the force argument is True and a specific version is
-    not requested then this function always checks for the latest version.
+    file is less than the measures_update_interval config value (days) before now then this function 
+    does nothing unless force is True. When this function checks for a more recent version and finds 
+    that the installed version is the most recent one then the modification time of that text file 
+    is changed to the current time even though nothing has changed in path. This limits the 
+    number of attempts to update the measures data (including checking
+    for more recent data) to once every measures_update_interval days. When the force argument 
+    is True and a specific version is not requested then this function always checks for the 
+    latest version. The measures_update_interval is always used as an int type (including
+    any truncation of the actual value in config if not an integer).
+
+    The measures_site is a single URL or a list of URLs to use to find the measures tar file
+    to use in the update. See measures_available for more details on how that parameter
+    is used.
+
+    If a specific version is requested then that must match a file in the list returned
+    by measures_available. The version is usually unique at each measures_site and so care
+    may need to be taken when requesting a specific version to ensure that it's available at 
+    measures_site. If measures_site is a list and a specific version is requested then
+    measures_update will try and find that version in each measures_site in that list, using
+    the first site that has that version.
 
     When auto_update_rules is True then path must exist and contain the expected readme.txt file.
     Path must be owned by the user, force must be False, and the version must be None. This 
@@ -107,12 +122,13 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
 
     Parameters
        - path (str=None) - Folder path to place updated measures data. Must contain a valid geodetic/readme.txt. If not set then config.measurespath is used.
-       - version (str=None) - Version of measures data to retrieve (usually in the form of WSRT_Measures_yyyymmdd-160001.ztar, see measures_available()). Default None retrieves the latest.
+       - version (str=None) - Version of measures data to retrieve (usually in the form of WSRT_Measures_yyyymmdd-160001.ztar, see measures_available). Default None retrieves the latest.
        - force (bool=False) - If True, always re-download the measures data. Default False will not download measures data if updated within the past day unless the version parameter is specified and different from what was last downloaded.
+       - measures_site(str or list of str = None) - Each value is a URL where measures tar files are found. If measures_site is a list then the elements are used in order until a list can be assembled. Default None uses config.measures_site.
        - logger (casatools.logsink=None) - Instance of the casalogger to use for writing messages. Default None writes messages to the terminal
        - auto_update_rules (bool=False) - If True then the user must be the owner of path, version must be None, and force must be False.
        - use_astron_obs_table (bool=False) - If True and force is also True then keep the Observatories table found in the Measures tar tarball (possibly overwriting the Observatories table from casarundata).
-       - verbose (int=None) - Level of output, 0 is none, 1 is to logger, 2 is to logger and terminal, defaults to casaconfig_verbose in config dictionary.
+       - verbose (int=None) - Level of output, 0 is none, 1 is to logger, 2 is to logger and terminal, defaults to config.casaconfig_verbose.
         
     Returns
        None
@@ -130,16 +146,11 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
     
     """
     import os
-    import pkg_resources
     from datetime import datetime
     import sys
 
     import tarfile
     import re
-    import ssl
-    import urllib.request
-    import certifi
-    import shutil
 
     from casaconfig import measures_available
     from casaconfig import AutoUpdatesNotAllowed, UnsetMeasurespath, RemoteError, NotWritable, BadReadme, BadLock, NoReadme, NoNetwork
@@ -148,16 +159,20 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
     from .get_data_lock import get_data_lock
     from .get_data_info import get_data_info
     from .measures_available import measures_available
-    
+    from .do_untar_url import do_untar_url
+
+    from .. import config as _config
+
     if path is None:
-        from .. import config as _config
         path = _config.measurespath
 
     if path is None:
         raise UnsetMeasurespath('measures_update: path is None and has not been set in config.measurespath. Provide a valid path and retry.')
 
+    if measures_site is None:
+        measures_site = _config.measures_site
+
     if verbose is None:
-        from .. import config as _config
         verbose = _config.casaconfig_verbose
 
     path = os.path.expanduser(path)
@@ -176,42 +191,52 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
         # make dirs all the way down, if possible
         os.makedirs(path)
         
-    current = None
+    currentVersion = None
     ageRecent = False
+    currentSite = None
 
     # first, does this look like it needs to be updated
 
     # get any existing measures readme information
     readmeInfo = get_data_info(path, logger, type='measures')
     if readmeInfo is not None:
-        current = readmeInfo['version']
+        currentVersion = readmeInfo['version']
+        currentSite = readmeInfo['site']
         if readmeInfo['age'] is not None:
-            ageRecent = readmeInfo['age'] < 1.0
+            ageRecent = readmeInfo['age'] < int(_config.measures_update_interval)
 
     if not force:
-        # don't check for new version if the age is less than 1 day
+        # don't check for new version if the age is less than measures_update_interval days
         if version is None and ageRecent:
-            if verbose > 0:
-                print_log_messages('measures_update: version installed or checked less than 1 day ago, nothing updated or checked', logger, verbose=verbose)
+            print_log_messages('measures_update: version installed or checked less than %s day(s) ago, nothing updated or checked' % int(_config.measures_update_interval), logger, verbose=verbose)
             return
         
-        # don't overwrite something that looks bad unless forced to do so
-        if current == 'invalid':
-            raise NoReadme('measures_update: no measures readme.txt file found at %s. Nothing updated or checked.' % path)
-        
-        if current == 'error':
-            raise BadReadme('measures_update: the measures readme.txt file at %s could not be read as expected, an update can not proceed unless force is True' % path)
-
-        # don't overwrite something that looks like valid measures data unless forced to do so
-        if current == 'unknown':
-            print_log_messages('measures_update: the measures data at %s is not maintained by casaconfig and so it can not be updated unless force is True' % path, logger)
-            return
+        if currentVersion is not None:
+            # don't overwrite something that looks bad unless forced to do so
+            if 'invalid' in currentVersion:
+                raise NoReadme('measures_update: no measures readme.txt file found at %s. Nothing updated or checked.' % path)
+
+            if 'error' in currentVersion:
+                msg = "measures_update: the measures readme.txt file at %s could not be read as expected, an update can not proceed unless force is True" % path
+                # add any additional error message, anything after the first ":" if is found
+                colonIndex = currentVersion.find(":")
+                if colonIndex != -1:
+                    msg += "; " + currentVersion[colonIndex+1:]
+                raise BadReadme(msg)
+
+            # don't overwrite something that looks like valid measures data unless forced to do so
+            if 'unknown' in currentVersion:
+                print_log_messages('measures_update: the measures data at %s is not maintained by casaconfig and so it can not be updated unless force is True' % path, logger, True)
+                return
 
         checkVersion = version
         if checkVersion is None:
             # get the current most recent version
             try:
-                checkVersion = measures_available()[-1]
+                # if measures_site is a list, measures_available will return the list from the appropriate site in that list
+                # that might not be the same site recorded with currentVersion, but that's OK for this update step
+                # the currentVersion is already old enough to trigger this check for a more recent version
+                checkVersion = measures_available(measures_site=measures_site)[-1]
             except NoNetwork as exc:
                 # no network, no point in continuing, just reraise
                 raise exc
@@ -220,12 +245,14 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
                 raise exc                
             except:
                 # unsure what happened, leave it at none, which will trigger an update attempt, which might work
+                print("unexpected excption in measuress_available")
+                import traceback
+                traceback.print_exc()
                 pass
 
         # don't re-download the same data
-        if (checkVersion is not None) and (checkVersion == current):
-            if verbose > 0:
-                print_log_messages('measures_update: requested version already installed in %s' % path, logger, verbose=verbose)
+        if (checkVersion is not None) and (checkVersion == currentVersion):
+            print_log_messages('measures_update: requested version already installed in %s' % path, logger, verbose=verbose)
             # update the age of the readme to now
             readme_path = os.path.join(path,'geodetic/readme.txt')
             # readme_path should already exist if it's here
@@ -239,7 +266,7 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
             msgs = []
             msgs.append("Error: the Observatories table was not found as expected in %s" % path)
             msgs.append("Either install casarundata first or set use_astron_obs_table and force to be True when using measures_update.")
-            msgs.append("Note that the Observatories table provided in the Astron measures tarfile is not the same as that maintained by CASA")
+            msgs.append("Note that the Observatories table provided in the measures tarfile is not the same as that maintained by CASA")
             print_log_messages(msgs, logger, True)
             return
         
@@ -253,7 +280,7 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
     lock_fd = None
     clean_lock = True    # set to false if the contents are actively being update and the lock file should not be cleaned one exception
     try:
-        print_log_messages('measures_update ... acquiring the lock ... ', logger)
+        print_log_messages('measures_update ... acquiring the lock ... ', logger, verbose=verbose)
 
         # the BadLock exception that may happen here is caught below
         lock_fd = get_data_lock(path, 'measures_update')
@@ -262,21 +289,19 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
         
         if not do_update:
             # recheck the readme file, another update may have already happened before the lock was obtained
-            current = None
+            currentVersion = None
             ageRecent = False
             
             readmeInfo = get_data_info(path, logger, type='measures')
             if readmeInfo is not None:
-                current = readmeInfo['version']
+                currentVersion = readmeInfo['version']
                 if readmeInfo['age'] is not None:
-                    ageRecent = readmeInfo['age'] < 1.0
+                    ageRecent = readmeInfo['age'] < int(_config.measures_update_interval)
 
-            if (version is not None) and (version == current):
-                # no update will be done, version is as requested - always verbose here because the lock is in use
-                print_log_messages('The requested measures version is already installed in %s, using version %s' % (path, current), logger)
+            if (version is not None) and (version == currentVersion):
+                print_log_messages('The requested measures version is already installed in %s, using version %s' % (path, currentVersion), logger, verbose=verbose)
             elif (version is None) and ageRecent:
-                # no update will be done, it's already been checked or updated recently - always verbose here because the lock is in use
-                print_log_messages('The latest measures version was checked recently in %s, using version %s' % (path, current), logger)
+                print_log_messages('The latest measures version was checked recently in %s, using version %s' % (path, currentVersion), logger, verbose=verbose)
             else:
                 # final check for problems before updating
                 if not force and readmeInfo is not None and (version=='invalid' or version=='unknown'):
@@ -291,39 +316,62 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
 
         if do_update:
             if force:
-                print_log_messages('A measures update has been requested by the force argument', logger)
-
-            print_log_messages('  ... finding available measures at www.astron.nl ...', logger)
-
-            files = measures_available()
-
-            # target filename to download
-            # for the non-force unspecified version case this can only get here if the age is > 1 day so there should be a newer version
-            # but that isn't checked - this could install a version that's already installed
-            target = files[-1] if version is None else version
-            if target not in files:
-                print_log_messages("measures_update can't find specified version %s" % target, logger, True)
-
+                print_log_messages('A measures update has been requested by the force argument', logger, verbose=verbose)
+
+            print_log_messages('  ... finding available measures ...', logger, verbose=verbose)
+
+            target = None
+            site = None
+
+            # if a specific version has been request AND measures_site is a list with more than one value
+            # then find the site with that version in it. Turn off logging in measures_available calls
+            # here as the age of the site doesn't matter
+            if version is not None and isinstance(measures_site, list) and len(measures_site) > 1:
+                target = version
+                for this_site in measures_site:
+                    files = measures_available(measures_site=this_site, logger=None)
+                    # make sure that site is excluded from files before target is checked
+                    files = files[1:]
+                    if target in files:
+                        site = this_site
+                        break
+                if site is None:
+                    print_log_messages("measures_update can't find specfied version %s at any of the sites in the measures_site list" % target, logger, True)
+                # else site and  target are set
             else:
+                # if measure_available sorts out what site to use if it's a list with more than one element
+                # a specific version probably needs to use a specific site to make it work
+                # if a specific version has been requested then the age of the site isn't an issue, turn logging off
+                this_logger = logger if version is None else None
+                files = measures_available(measures_site=measures_site, logger=this_logger)
+                site = files[0]
+                # this makes sure that the site is exclude from files before target is checked.
+                files = files[1:]
+
+                # target filename to download
+                # for the non-force unspecified version case this can only get here if the age is > 1 day so there should be a newer version
+                # but that isn't checked - this could install a version that's already installed
+                target = files[-1] if version is None else version
+                if target not in files:
+                    print_log_messages("measures_update can't find specified version %s at site %s" % (target,site), logger, True)
+                    # unset site here to signal nothing to extract
+                    site = None
+
+            if site is not None:
                 # there are files to extract
-                print_log_messages('  ... downloading %s from ASTRON server to %s ...' % (target, path), logger)
+                print_log_messages('  ... downloading %s from %s to %s ...' % (target, site, path), logger, verbose=verbose)
 
-                astronURL = 'https://www.astron.nl/iers'
-                context = ssl.create_default_context(cafile=certifi.where())
-                # just in case there's a redirect at astron the way there is for the go.nrao.edu site and casarundata
-                measuresURLroot = urllib.request.urlopen(astronURL, context=context).url
-                measuresURL = os.path.join(measuresURLroot, target)
-                
                 # it's at this point that this code starts modifying what's there so the lock file should
-                # not be removed on failure after this although it may leave that temp tar file around, but that's OK
+                # not be removed on failure from here on (until it succeeds)
                 clean_lock = False
+                
                 # remove any existing measures readme.txt now in case something goes wrong during extraction
                 readme_path = os.path.join(path,'geodetic/readme.txt')
                 if os.path.exists(readme_path):
                     os.remove(readme_path)
 
-                # custom filter that incorporates data_filter to watch for dangerous members of the tar file and
-                # add filtering to remove the Observatories table (unless use_astron_obs_table is True) and
+                # custom filter that incorporates 'data_filter' to watch for dangerous members of the tar file
+                # this adds filtering to remove the Observatories table (unless use_astron_obs_table is True) and
                 # the *.old tables that may be in the geodetic tree
                 def custom_filter(member, extractionPath):
                     # member is a TarInfo instance and extractionPath is the destination path
@@ -337,18 +385,16 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
                         member = None
                     return member
 
-                with urllib.request.urlopen(measuresURL, context=context, timeout=400) as tstream, tarfile.open(fileobj=tstream, mode='r|*') as tar :
-                    # use the 'data' filter if available, revert to previous 'fully_trusted' behavior of not available
-                    tar.extraction_filter = custom_filter
-                    tar.extractall(path=path)
-                    tar.close()
+                # untar the target at site to path using the custom filter, do not be verbose
+                do_untar_url(site, target, path, custom_filter)
 
                 # create a new readme.txt file
                 with open(readme_path,'w') as fid:
-                    fid.write("# measures data populated by casaconfig\nversion : %s\ndate : %s" % (target, datetime.today().strftime('%Y-%m-%d')))
+                    # site added last to not break past readers of the measures readme
+                    fid.write("# measures data populated by casaconfig\nversion : %s\ndate : %s\nsite : %s" % (target, datetime.today().strftime('%Y-%m-%d'),site))
 
                 clean_lock = True
-                print_log_messages('  ... measures data updated at %s' % path, logger)
+                print_log_messages('  ... measures data updated at %s' % path, logger, verbose=verbose)
 
             # closing out the do_update
 
@@ -372,6 +418,14 @@ def measures_update(path=None, version=None, force=False, logger=None, auto_upda
         msgs.append('Check for other updates in progress or choose a different path or clear out this path and reinstall the casarundata as well as the measures data')
         print_log_messages(msgs, logger, True)
         raise
+
+    except RemoteError as exc:
+        # probably an empty site
+        msgs = [str(exc)]
+        msgs.append('There was a remote error while attempting to update measures at %s' % path)
+        msgs.append('Check the value of measures_site and try again')
+        print_log_messages(msgs, logger, True)
+        raise
         
     except Exception as exc:
         msgs = []

casaconfig/private/print_log_messages.py

@@ -30,13 +30,17 @@ def print_log_messages(msg, logger, is_err=False, verbose=2):
        - msg (str) - The message to print and optionally log.
        - logger (casatools.logsink) - Instance of the casalogger to use. Not used if None.
        - is_err (bool=False) - When False, output goes to sys.stdout and logged as INFO level. When True, output goes to sys.stderr and logged as SEVERE
-       - verbose (int=2) - When < 2 then msg is only printed if there is no logger, otherwise it's just logged
+       - verbose (int=2) - 0 - do nothing, 1 log unless there is no logger, then print, 2 log and print
 
     Returns
        None
 
     """
 
+    if verbose == 0:
+        # do nothing
+        return
+
     import sys
     fileout = sys.stdout
     loglevel = 'INFO'
@@ -48,8 +52,8 @@ def print_log_messages(msg, logger, is_err=False, verbose=2):
     if type(msg) is not list:
         msg = [msg]
 
-    # always print if there is no logger, if there is a logger and verbose is < 2 then do not print
-    if (logger is None) or (not verbose < 2):
+    # print if there is no logger or verbose >= 2
+    if (logger is None) or (verbose >= 2):
         for m_msg in msg:
             print(m_msg,file=fileout)
 

casaconfig/private/pull_data.py

@@ -22,6 +22,7 @@ def pull_data(path=None, version=None, force=False, logger=None, verbose=None):
     The verbose argument controls the level of information provided when this function when the data
     are unchanged for expected reasons. A level of 0 prints and logs nothing. A
     value of 1 logs the information and a value of 2 logs and prints the information.
+    Error messages are always logged and printed.
 
     The path must either contain a previously installed version of casarundata
     or it must not exist or be empty.
@@ -107,19 +108,23 @@ def pull_data(path=None, version=None, force=False, logger=None, verbose=None):
 
     from casaconfig import data_available
     from casaconfig import get_data_info
-    from casaconfig import UnsetMeasurespath, BadLock, BadReadme, NotWritable
+    from casaconfig import UnsetMeasurespath, BadLock, BadReadme, NoNetwork, NotWritable, RemoteError
 
     from .print_log_messages import print_log_messages
     from .get_data_lock import get_data_lock
     from .do_pull_data import do_pull_data
 
+    from .. import config as _config
+
     if path is None:
-        from .. import config as _config
         path = _config.measurespath
 
     if path is None:
         raise UnsetMeasurespath('path is None and has not been set in config.measurespath. Provide a valid path and retry.')
 
+    if verbose is None:
+        verbose = _config.casaconfig_verbose
+        
     # when a specific version is requested then the measures readme.txt that is part of that version
     # will get a timestamp of now so that default measures updates won't happen for a day unless the
     # force argument is used for measures_update
@@ -223,7 +228,7 @@ def pull_data(path=None, version=None, force=False, logger=None, verbose=None):
     lock_fd = None
     clean_lock = True  # set to False if the contents are actively being update and the lock file should not be cleaned on exception
     try:
-        print_log_messages('pull_data using version %s, acquiring the lock ... ' % version, logger)
+        print_log_messages('pull_data using version %s, acquiring the lock ... ' % version, logger, verbose=verbose)
 
         # attempting to get the lock will raise NoNetwork if there is no network and the lock will not be set, catch and reemit that in this try block
         lock_fd = get_data_lock(path, 'pull_data')
@@ -248,13 +253,13 @@ def pull_data(path=None, version=None, force=False, logger=None, verbose=None):
                             if measuresVersion == expectedMeasuresVersion:
                                 # no pull is necessary
                                 do_pull = False
-                                print_log_messages('pull_data requested "release" versions of casarundata and measures are already installed.', logger)
+                                print_log_messages('pull_data requested "release" versions of casarundata and measures are already installed.', logger, verbose=verbose)
                             # otherwise this is a release pull and even if the measuresVersion is 'unknown' or 'invalid' we should proceed with a pull at this point
                             # problems with casarundata path will have been caught before
                     else:
                         # nothing to do here, already at the expected casarundata version and a pull is not being forced and this is not a release pull
                         do_pull = False
-                        print_log_messages('pull_data requested version is already installed.', logger)
+                        print_log_messages('pull_data requested version is already installed.', logger, verbose=verbose)
 
                 # a version of 'invalid', 'error', or 'unknown' is a surprise here, likely caused by something else doing something
                 # incompatible with this attempt
@@ -274,7 +279,7 @@ def pull_data(path=None, version=None, force=False, logger=None, verbose=None):
         if do_pull:
             # do not clean the lock file contents at this point unless do_pull_data returns normally
             clean_lock = False
-            do_pull_data(path, version, installed_files, currentVersion, currentDate, logger)
+            do_pull_data(path, version, installed_files, currentVersion, currentDate, logger, verbose)
             clean_lock = True
             if namedVersion:
                 # a specific version has been requested, set the times on the measures readme.txt to now to avoid

casaconfig/private/read_readme.py

@@ -16,45 +16,89 @@ def read_readme(path):
     """
     Read and parse the contents of a readme.txt file used by casaconfig.
 
-    A dictionary containing the 'version', 'date', and 'extra' (containing
+    A dictionary containing the 'version', 'date', 'site', and 'extra' (containing
     a list of optional extra lines found). On error, the return values is None.
 
     The extra lines are stripped and do not include lines begining with '#'
 
     The format is assumed to be:
         a line begining with #, which is ignored.
+        a line "site: the site url string"
         a line "version : the versions string"
         a line "date : the date"
         optional extra lines (the manifest of installed files for the main readme)
 
-    The version string and date are stripped of leading and trailing whitespace.
+    The version, site and date strings are stripped of leading and trailing whitespace.
+
+    The site string only appears in measures data and is missing is early readme
+    files for measures data. If no site string is seen the value in the dictionary
+    is None.
+
+    The version and date strings are required. If they are not found a BadReadme
+    exception is raised.
 
     Parameters
        - path (str) - the path to the file to be read
 
     Returns
-       Dictionary of 'version' (the version string), 'date' (the date string),
-             'extra' (a list of any extra lines found). The return value is None on error.
+       Dictionary of 'site' ( the site URL or None), 'version' (the version string), 
+             'date' (the date string), 'extra' (a list of any extra lines found). 
+             The return value is None on error except that format errors raise
+             a BadReadme exception.
+
+    Raises
+       - casaconfig.BadReadme - raised when there is a format error in the file at path
     """
 
     import os
+    from casaconfig import BadReadme
 
+    site = None
     version = ""
     date = ""
     extra = []
     result = None
-    
+
     try:
         with open(path, 'r') as fid:
             readmeLines = fid.readlines()
-            version = readmeLines[1].split(':')[1].strip()
-            date = readmeLines[2].split(':')[1].strip()
-            if len(readmeLines) > 3:
-                for extraLine in readmeLines[3:]:
-                    if (extraLine[0] != '#'):
-                        extra.append(extraLine.strip())
-            result = {'version':version, 'date':date, 'extra':extra}
-    except:
+            # order is unimportant except the first line must start with
+            # an "#" and the second line starting with "#" indicates the
+            # start of the "extra" lines
+            # duplicate instances of the same key in key : value lines are accepted
+            # the last value associated with key is used, that should never happen.
+            inExtraLines = False
+            if readmeLines[0][0] != '#':
+                raise BadReadme('Readme file missing required "#" at start of first line, %s' % path)
+            for line in readmeLines[1:]:
+                if inExtraLines and not (line[0] == '#'):
+                    extra.append(line.strip())
+                elif line[0] == '#':
+                    inExtraLines = True
+                else:
+                    splitLine = line.split(':')
+                    if len(splitLine) < 2:
+                        raise BadReadme('A line did not have an expected ":" separating into at least two parts, %s' % path)
+                    key = splitLine[0].strip()
+                    if key == 'site':
+                        # the URL likely also has a ":" so rejoin the second part and then strip that
+                        siteURL = ":".join(splitLine[1:])
+                        site = siteURL.strip()
+                    elif key == 'version':
+                        version = splitLine[1].strip()
+                    elif key == 'date':
+                        date = splitLine[1].strip()
+                    # anything else is OK and silently ignored
+        result = {'site':site, 'version':version, 'date':date, 'extra':extra}
+    except BadReadme as exc:
+        # reraise it
+        raise
+    except Exception as exc:
+        # silently return None, handled elsewhere
         result = None
 
+    # require that date and version are now set in result
+    if result is not None and ((len(result['version'])==0) or (len(result['date'])==0)):
+        raise BadReadme('Readme file missing required version or date fields, %s' % path)
+
     return result

casaconfig/private/set_casacore_path.py

@@ -32,12 +32,12 @@ def set_casacore_path(path=None):
        None
 
     """
-    import pkg_resources
+    import importlib.resources
     import os
     import re
     import sys
     
-    if path is None: path = pkg_resources.resource_filename('casaconfig', 'data/')
+    if path is None: path = str(importlib.resources.files('casaconfig').joinpath('data/'))
     path = os.path.abspath(os.path.expanduser(path))
 
     rctext = 'measures.directory: %s\n' % path

casaconfig/private/summary.py

@@ -80,6 +80,9 @@ def summary(configDict = None):
             msg += "; legacy data not maintained by casaconfig"
         elif measVers == "error":
             msg += "; unexpected readme.txt file, measures should be reinstalled"
+        measSite = dataInfo['measures']['site']
+        if measSite is not None and len(measSite) > 0:
+            msg = msg + " site : " + measSite
         print(msg)
     
     if (dataInfo['release'] is None):