Compatibility with NegMAS 0.11.2

Avoid pypi test

Adding ANLNegotiator to support propose/respond. Negotiators for SAO have the following interface: __call__() which is the default for SAONegotiator. This is implemented by calling propose() and respond() whenever needed. ANL agents implemented __call__ directly. The ANLNegotiator provides an implementation of propose() and respond() assuming __call__ is called directly.

Renaming old builtin file

Adding –complete-only/allow-parital to CLI. Controls whether partially run tournaments are considered

Show debug errors in cli

Upgrading GitHub Actions to avoid node 16

Adding tournament display and combination to CLI

Adding –sort-runs to the CLI

Passing hidden_time_limit to the Cartesian tournament

Avoiding plotting issue in windows (not fully tested)

Requiring latest negmas (v0.10.23)

fixing issue #1 (anlv failure on macOS)

Correcting git installation method in docs

requiring negmas 0.10.21

Better comments on micro

Defaulting to 3min hidden time limit per negotiation

pypi workflow update

Fix tutorial links

Adding divide-the-pies scenarios

Adding workflow to test on negmas master

Tutorial and docs update

Update faq

bugfix in visualizer initial tournament list

Correcting auto pushing to PyPi

Adding simple dockerfile

Adding –port, –address to anlv show. You can now set the port and address of the visualizer

Visualizer parses folders recursively

minor: faster saving of figs

Adding mkdocs to dev requirements

Removing NaiveTitForTat from the default set of competitors

Improving tutorial

Improved visualizer

• Adding filtering by scenario or strategy to the main view.

• Adding new options to show scenario statistics, scenario x strategy statistics, and cases with no agreements at all.

• You can show multiple negotiations together

• You can show the descriptive statistics of any metric according to strategy or scenario

• More plotting options for metrics

Improved CLI

• Adding the ability to pass parameters to competitors in the CLI.

• Removing NaiveTitForTat from the default set of competitors

• Making small tournaments even smaller

New and improved strategies

• Adding RVFitter strategy which showcases simple implementation of curve fitting for reserved value estimation and using logging.

• Adding more comments to NashSeeker strategy

• Simplified implementation of MiCRO

• Adding a simple test for MiCRO

• Avoid failure when Nash cannot be found in NashSeeker

Migrating to NegMAS 0.10.11. Needed for logging (and 0.10.10 is needed for self.oppponent_ufun)

Changing default order of agents

Adding a basic visualizer

Adding make-scenarios to the CLI

Passing opponent ufun in the private info

Separating implementation of builtin agents

requiring NegMAS 0.10.9

Retrying scenario generation if it failed

Defaulting to no plotting in windows

Defaulting to no-plotting on windows to avoid an error caused by tkinter

Retry scenario generation on failure. This is useful for piece-wise linear which will fail (by design) if n_pareto happened to be less than n_segments + 1

Adding better scenario generation and supporting mixtures of zero-sum, monotonic and general scenarios.

Requiring negmas 0.10.8

Controlling log path in anl2024_tournament() through the added base_path argument

Added anl cli for running tournaments.

Added the ability to hide or show type names during negotiations

Corrected a bug in importing unique_name

Now requires negmas 0.10.6

Adding ANL 2024 placeholder