Commit Graph

111 Commits

Author SHA1 Message Date
Sarah Hoffmann
7158433cd3 disable warning about non-toplevel import
They are needed here so nominatim can be run when osmium
is not installed. Everything except replication will work fine.
2021-01-30 16:29:28 +01:00
Sarah Hoffmann
e629a175ed introduce custom UsageError
This is a exception to be thrown when the error occures because
of bad user data. We don't want to print a full stack trace in
these cases but just tell the user what went wrong.
2021-01-30 16:20:10 +01:00
Sarah Hoffmann
4cb6dc01f3 port replication update function to python 2021-01-30 15:50:34 +01:00
Sarah Hoffmann
8f0885f6cb port check-for-update function to python 2021-01-28 14:50:14 +01:00
Sarah Hoffmann
d78f0ba804 port replication initialisation to Python 2021-01-26 22:50:54 +01:00
Sarah Hoffmann
5b46fcad8e convert functon creation to python
The new functions always creates normal and partitioned functions.
Also adds specialised connection and cursor classes for adding
frequently used helper functions.
2021-01-26 22:50:54 +01:00
Sarah Hoffmann
94fa7162be port address level computation to Python
Also adds simple tests for correct table creation.
2021-01-26 22:50:54 +01:00
Sarah Hoffmann
e6c2842b66 move update code for postcode and word count to Python
Adds also tests for the new function to execute a SQL script.
2021-01-26 22:50:54 +01:00
Sarah Hoffmann
e6d9485c4a cli: import python modules for commands on demand
Given that only one command will be executed in the end, it is
not necessary to import what amounts to the whole library. This
becomes in particular important for update functions that have
a dependency on pyosmium. The dependency can remain optional for
people not using updates.
2021-01-26 22:50:54 +01:00
Sarah Hoffmann
42ec67f63c add more tests for CLI parameter parser 2021-01-20 21:30:27 +01:00
Sarah Hoffmann
8c02786820 add tests for indexer 2021-01-20 21:30:27 +01:00
Sarah Hoffmann
c26f323bf5 add simple tests for CLI parsing 2021-01-20 21:30:27 +01:00
Sarah Hoffmann
041ae67fd9 optionally hand in command line arguments to CLI functions
Allows easier testing.
2021-01-20 21:30:27 +01:00
Sarah Hoffmann
52b76d1d01 add tests for Python exec_utils 2021-01-20 21:30:27 +01:00
marc tobias
f62c784102 correct parameter name in query CLI 2021-01-20 21:09:41 +01:00
Sarah Hoffmann
8cf54a1317 add API functions to nominatim tool 2021-01-19 19:38:46 +01:00
Sarah Hoffmann
77e287f669 rename nominatim.admin to nominatim.tools 2021-01-19 19:38:46 +01:00
Sarah Hoffmann
5d95a72758 probe for php_cgi in cmake to be used for querying 2021-01-19 19:38:46 +01:00
Sarah Hoffmann
504922ffbe remove old nominatim.py in favour of 'nominatim index'
The PHP scripts need to know the position of the nominatim
tool in order to call it. This is handed in as environment
variable, so it can be set by the Python script.
2021-01-18 15:43:27 +01:00
Sarah Hoffmann
c77877a934 implementaion of 'nominatim index' 2021-01-18 15:43:27 +01:00
Sarah Hoffmann
27977411e9 move indexing function into its own Python module
This makes it mow a standard function of our new Python
library instead of a stand-alone program.
2021-01-18 15:43:27 +01:00
Sarah Hoffmann
b79c79fa73 add function to get a DSN for psycopg
Converts the PHP DSN syntax into psycopg syntax when necessary.
2021-01-18 15:43:27 +01:00
Sarah Hoffmann
7cf9d459d6 use check parameter of subprocess.run
...instead of checking on our own.

Also increase required version of Python to 3.5 because of
subprocess.run().
2021-01-15 10:43:04 +01:00
Sarah Hoffmann
8e53f63036 fix errors reported by pylint 2021-01-15 08:57:00 +01:00
Sarah Hoffmann
eda0900c8e fix typo 2021-01-14 20:30:27 +01:00
Sarah Hoffmann
2f73bb3643 bdd: directly call utility scripts in lib
This removes the dependency on php-symfony-dotenv for the tests.
2021-01-14 18:19:22 +01:00
Sarah Hoffmann
88c57b4dc8 maller command execution fixes 2021-01-14 12:03:49 +01:00
Sarah Hoffmann
ba13cfd9ff make sure that environment variables have highest prio 2021-01-14 11:12:45 +01:00
Sarah Hoffmann
1ff8751caa liniting of new python code 2021-01-14 10:19:21 +01:00
Sarah Hoffmann
98dbc84836 add wrapper calls for all nominatim tool functions 2021-01-14 09:37:47 +01:00
Sarah Hoffmann
04690ad8c4 implement warming in new cli tool
Adds infrastructure for calling the legacy PHP scripts. As the
CONST_* values cannot be set from the python script, hand the values
in via secret environment variables instead. These are all
temporary hacks for the transition phase to python code.
2021-01-13 18:25:15 +01:00
Sarah Hoffmann
d6bcb7c8b7 consolidate cli interface to single tool 2021-01-13 10:11:58 +01:00
Sarah Hoffmann
57f5e6d898 create skeleton for new CLI tools 2021-01-12 22:21:20 +01:00
Sarah Hoffmann
612fd50612 add skeleton for new Nominatim executables 2021-01-12 10:17:28 +01:00
Sarah Hoffmann
5016eace34 improve progress logging during indexing
Wait for 2 seconds before logging the first progress, so that we
have numbers that are a bit more reliable statistically speaking.

Also provides an actual implementation for the log_interval
parameter and fixes some small style issues.
2020-11-30 10:59:29 +01:00
Sarah Hoffmann
4ac29fb525 only index larger batches for rank 30
Fixes #2045.
2020-11-05 22:14:49 +01:00
Sarah Hoffmann
13dba94307 do not run rank 0 objects in parallel
Waterways are at address rank 0 and do linking. This might lead to
deadlocks.
2020-08-22 19:51:19 +02:00
Sarah Hoffmann
73c449b97b switch indexind to address rank
A place needs all lower address rank object indexed to make up
the address. The search rank no longer ensures that as it can have
a different ordering than the address rank.

This switches indexing rank order to address ranks. Non-address
objects (with address rank 0) are indexed together with POIs.
2020-08-18 16:58:58 +02:00
Sarah Hoffmann
3816b86a9e nominatim: also index boundaries by rank
We need to make sure that the entry in serach_name from a lower rank
is indeed available.
2020-08-18 15:17:09 +02:00
Sarah Hoffmann
a4b30fc649 index admin boundaries before everything else
Avoids irregularities that might happen because the address
rank of a boundary is changed through linking.
2020-08-18 15:17:09 +02:00
Sarah Hoffmann
fc50eb8688 nominatim: move DBConnection class into its own file 2020-08-18 15:17:09 +02:00
Sarah Hoffmann
5be084e0f5 indexer: allow batch processing of places
Request and process multiple place_ids at once so that
Postgres can make better use of caching and there are less
transactions running.
2020-08-03 10:32:39 +02:00
Sarah Hoffmann
2323923bec indexer: move progress tracker into separate class 2020-08-03 10:32:39 +02:00
Sarah Hoffmann
0f54d42863 indexer: get rid of special handling of few places
Given that we do not distiribute geometry sectors to threads anymore,
there is no point in this kind of special handling.
2020-08-03 10:32:39 +02:00
Sarah Hoffmann
cca366196d Disable JIT and parallel workers when indexing
Locally disable jit and parallel workers in the connection that
do indexing. The query planner tends to be overenthusiatic about
using JIT. But with the rather less complex queries we have, the
overhead tends to be larger than the performance gain.

Fixes #1677.
2020-05-30 11:20:16 +02:00
Sarah Hoffmann
431948d768 nominatim: always use deadlock-protected wait
Fixes #1785.
2020-05-15 18:49:27 +02:00
Sarah Hoffmann
5469d02d03 nominatim.py: fix wrong use of assert
Fixes #1762.
2020-04-19 17:59:49 +02:00
Sarah Hoffmann
d1eeaa59a6 nominatim.py: use async in connect() function
The _async parameter name is only supported since psycopg 2.7.
However, async is a keyword in Python >= 3.7, so using this
gives us a syntax error. Working around this by defining the
parameters in a dict and handing that into the connect function.
2020-02-11 22:16:17 +01:00
Sarah Hoffmann
882f496e0a nominatim.py: also catch deadlocks on final wait 2020-02-11 22:16:17 +01:00
Sarah Hoffmann
8b8aa1b4e6 regularly close connection while indexing
Postgres sooner or later runs out of memory when the connection
is used for too long.
2020-02-11 22:16:17 +01:00